community.general/lib/ansible/parsing/vault/__init__.py

1333 lines
51 KiB
Python
Raw Normal View History

# (c) 2014, James Tanner <tanner.jc@gmail.com>
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
# (c) 2016, Adrian Likins <alikins@redhat.com>
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
# (c) 2016 Toshio Kuratomi <tkuratomi@ansible.com>
#
# Ansible is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# Ansible is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with Ansible. If not, see <http://www.gnu.org/licenses/>.
2014-10-29 00:44:21 +00:00
# Make coding more python3-ish
from __future__ import (absolute_import, division, print_function)
__metaclass__ = type
import os
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
import random
import shlex
import shutil
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
import subprocess
import sys
import tempfile
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
import warnings
from binascii import hexlify
from binascii import unhexlify
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
from binascii import Error as BinasciiError
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
HAS_CRYPTOGRAPHY = False
HAS_PYCRYPTO = False
HAS_SOME_PYCRYPTO = False
CRYPTOGRAPHY_BACKEND = None
try:
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
with warnings.catch_warnings():
warnings.simplefilter("ignore", DeprecationWarning)
from cryptography.exceptions import InvalidSignature
from cryptography.hazmat.backends import default_backend
from cryptography.hazmat.primitives import hashes, padding
from cryptography.hazmat.primitives.hmac import HMAC
from cryptography.hazmat.primitives.kdf.pbkdf2 import PBKDF2HMAC
from cryptography.hazmat.primitives.ciphers import (
Cipher as C_Cipher, algorithms, modes
)
CRYPTOGRAPHY_BACKEND = default_backend()
HAS_CRYPTOGRAPHY = True
except ImportError:
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
pass
try:
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
from Crypto.Cipher import AES as AES_pycrypto
HAS_SOME_PYCRYPTO = True
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
# Note: Only used for loading obsolete VaultAES files. All files are written
# using the newer VaultAES256 which does not require md5
from Crypto.Hash import SHA256 as SHA256_pycrypto
from Crypto.Hash import HMAC as HMAC_pycrypto
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
# Counter import fails for 2.0.1, requires >= 2.6.1 from pip
from Crypto.Util import Counter as Counter_pycrypto
# KDF import fails for 2.0.1, requires >= 2.6.1 from pip
from Crypto.Protocol.KDF import PBKDF2 as PBKDF2_pycrypto
HAS_PYCRYPTO = True
except ImportError:
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
pass
from ansible.errors import AnsibleError, AnsibleAssertionError
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
from ansible import constants as C
from ansible.module_utils.six import PY3, binary_type
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
# Note: on py2, this zip is izip not the list based zip() builtin
from ansible.module_utils.six.moves import zip
from ansible.module_utils._text import to_bytes, to_text, to_native
from ansible.utils.display import Display
from ansible.utils.path import makedirs_safe
display = Display()
2015-08-24 22:49:55 +00:00
b_HEADER = b'$ANSIBLE_VAULT'
CIPHER_WHITELIST = frozenset((u'AES', u'AES256'))
CIPHER_WRITE_WHITELIST = frozenset((u'AES256',))
# See also CIPHER_MAPPING at the bottom of the file which maps cipher strings
# (used in VaultFile header) to a cipher class
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
NEED_CRYPTO_LIBRARY = "ansible-vault requires either the cryptography library (preferred) or"
if HAS_SOME_PYCRYPTO:
NEED_CRYPTO_LIBRARY += " a newer version of"
NEED_CRYPTO_LIBRARY += " pycrypto in order to function."
class AnsibleVaultError(AnsibleError):
pass
class AnsibleVaultPasswordError(AnsibleVaultError):
pass
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
class AnsibleVaultFormatError(AnsibleError):
pass
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
def is_encrypted(data):
""" Test if this is vault encrypted data blob
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
:arg data: a byte or text string to test whether it is recognized as vault
encrypted data
:returns: True if it is recognized. Otherwise, False.
"""
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
try:
# Make sure we have a byte string and that it only contains ascii
# bytes.
b_data = to_bytes(to_text(data, encoding='ascii', errors='strict', nonstring='strict'), encoding='ascii', errors='strict')
except (UnicodeError, TypeError):
# The vault format is pure ascii so if we failed to encode to bytes
# via ascii we know that this is not vault data.
# Similarly, if it's not a string, it's not vault data
return False
if b_data.startswith(b_HEADER):
return True
return False
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
def is_encrypted_file(file_obj, start_pos=0, count=-1):
"""Test if the contents of a file obj are a vault encrypted data blob.
:arg file_obj: A file object that will be read from.
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
:kwarg start_pos: A byte offset in the file to start reading the header
from. Defaults to 0, the beginning of the file.
:kwarg count: Read up to this number of bytes from the file to determine
if it looks like encrypted vault data. The default is -1, read to the
end of file.
:returns: True if the file looks like a vault file. Otherwise, False.
"""
# read the header and reset the file stream to where it started
current_position = file_obj.tell()
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
try:
file_obj.seek(start_pos)
return is_encrypted(file_obj.read(count))
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
finally:
file_obj.seek(current_position)
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
def _parse_vaulttext_envelope(b_vaulttext_envelope, default_vault_id=None):
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_tmpdata = b_vaulttext_envelope.splitlines()
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_tmpheader = b_tmpdata[0].strip().split(b';')
b_version = b_tmpheader[1].strip()
cipher_name = to_text(b_tmpheader[2].strip())
vault_id = default_vault_id
# Only attempt to find vault_id if the vault file is version 1.2 or newer
# if self.b_version == b'1.2':
if len(b_tmpheader) >= 4:
vault_id = to_text(b_tmpheader[3].strip())
b_ciphertext = b''.join(b_tmpdata[1:])
return b_ciphertext, b_version, cipher_name, vault_id
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
def parse_vaulttext_envelope(b_vaulttext_envelope, default_vault_id=None, filename=None):
"""Parse the vaulttext envelope
When data is saved, it has a header prepended and is formatted into 80
character lines. This method extracts the information from the header
and then removes the header and the inserted newlines. The string returned
is suitable for processing by the Cipher classes.
:arg b_vaulttext: byte str containing the data from a save file
:kwarg default_vault_id: The vault_id name to use if the vaulttext does not provide one.
:kwarg filename: The filename that the data came from. This is only
used to make better error messages in case the data cannot be
decrypted. This is optional.
:returns: A tuple of byte str of the vaulttext suitable to pass to parse_vaultext,
a byte str of the vault format version,
the name of the cipher used, and the vault_id.
:raises: AnsibleVaultFormatError: if the vaulttext_envelope format is invalid
"""
# used by decrypt
default_vault_id = default_vault_id or C.DEFAULT_VAULT_IDENTITY
try:
return _parse_vaulttext_envelope(b_vaulttext_envelope, default_vault_id)
except Exception as exc:
msg = "Vault envelope format error"
if filename:
msg += ' in %s' % (filename)
msg += ': %s' % exc
raise AnsibleVaultFormatError(msg)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def format_vaulttext_envelope(b_ciphertext, cipher_name, version=None, vault_id=None):
""" Add header and format to 80 columns
:arg b_ciphertext: the encrypted and hexlified data as a byte string
:arg cipher_name: unicode cipher name (for ex, u'AES256')
:arg version: unicode vault version (for ex, '1.2'). Optional ('1.1' is default)
:arg vault_id: unicode vault identifier. If provided, the version will be bumped to 1.2.
:returns: a byte str that should be dumped into a file. It's
formatted to 80 char columns and has the header prepended
"""
if not cipher_name:
raise AnsibleError("the cipher must be set before adding a header")
version = version or '1.1'
# If we specify a vault_id, use format version 1.2. For no vault_id, stick to 1.1
if vault_id and vault_id != u'default':
version = '1.2'
b_version = to_bytes(version, 'utf-8', errors='strict')
b_vault_id = to_bytes(vault_id, 'utf-8', errors='strict')
b_cipher_name = to_bytes(cipher_name, 'utf-8', errors='strict')
header_parts = [b_HEADER,
b_version,
b_cipher_name]
if b_version == b'1.2' and b_vault_id:
header_parts.append(b_vault_id)
header = b';'.join(header_parts)
b_vaulttext = [header]
b_vaulttext += [b_ciphertext[i:i + 80] for i in range(0, len(b_ciphertext), 80)]
b_vaulttext += [b'']
b_vaulttext = b'\n'.join(b_vaulttext)
return b_vaulttext
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
def _unhexlify(b_data):
try:
return unhexlify(b_data)
except (BinasciiError, TypeError) as exc:
raise AnsibleVaultFormatError('Vault format unhexlify error: %s' % exc)
def _parse_vaulttext(b_vaulttext):
b_vaulttext = _unhexlify(b_vaulttext)
b_salt, b_crypted_hmac, b_ciphertext = b_vaulttext.split(b"\n", 2)
b_salt = _unhexlify(b_salt)
b_ciphertext = _unhexlify(b_ciphertext)
return b_ciphertext, b_salt, b_crypted_hmac
def parse_vaulttext(b_vaulttext):
"""Parse the vaulttext
:arg b_vaulttext: byte str containing the vaulttext (ciphertext, salt, crypted_hmac)
:returns: A tuple of byte str of the ciphertext suitable for passing to a
Cipher class's decrypt() function, a byte str of the salt,
and a byte str of the crypted_hmac
:raises: AnsibleVaultFormatError: if the vaulttext format is invalid
"""
# SPLIT SALT, DIGEST, AND DATA
try:
return _parse_vaulttext(b_vaulttext)
except AnsibleVaultFormatError:
raise
except Exception as exc:
msg = "Vault vaulttext format error: %s" % exc
raise AnsibleVaultFormatError(msg)
def verify_secret_is_not_empty(secret, msg=None):
'''Check the secret against minimal requirements.
Raises: AnsibleVaultPasswordError if the password does not meet requirements.
Currently, only requirement is that the password is not None or an empty string.
'''
msg = msg or 'Invalid vault password was provided'
if not secret:
raise AnsibleVaultPasswordError(msg)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
class VaultSecret:
'''Opaque/abstract objects for a single vault secret. ie, a password or a key.'''
def __init__(self, _bytes=None):
# FIXME: ? that seems wrong... Unset etc?
self._bytes = _bytes
@property
def bytes(self):
'''The secret as a bytestring.
Sub classes that store text types will need to override to encode the text to bytes.
'''
return self._bytes
def load(self):
return self._bytes
class PromptVaultSecret(VaultSecret):
default_prompt_formats = ["Vault password (%s): "]
def __init__(self, _bytes=None, vault_id=None, prompt_formats=None):
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
super(PromptVaultSecret, self).__init__(_bytes=_bytes)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
self.vault_id = vault_id
if prompt_formats is None:
self.prompt_formats = self.default_prompt_formats
else:
self.prompt_formats = prompt_formats
@property
def bytes(self):
return self._bytes
def load(self):
self._bytes = self.ask_vault_passwords()
def ask_vault_passwords(self):
b_vault_passwords = []
for prompt_format in self.prompt_formats:
prompt = prompt_format % {'vault_id': self.vault_id}
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
try:
vault_pass = display.prompt(prompt, private=True)
except EOFError:
raise AnsibleVaultError('EOFError (ctrl-d) on prompt for (%s)' % self.vault_id)
verify_secret_is_not_empty(vault_pass)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_vault_pass = to_bytes(vault_pass, errors='strict', nonstring='simplerepr').strip()
b_vault_passwords.append(b_vault_pass)
# Make sure the passwords match by comparing them all to the first password
for b_vault_password in b_vault_passwords:
self.confirm(b_vault_passwords[0], b_vault_password)
if b_vault_passwords:
return b_vault_passwords[0]
return None
def confirm(self, b_vault_pass_1, b_vault_pass_2):
# enforce no newline chars at the end of passwords
if b_vault_pass_1 != b_vault_pass_2:
# FIXME: more specific exception
raise AnsibleError("Passwords do not match")
Vault secrets script client inc new 'keyring' client (#27669) This adds a new type of vault-password script (a 'client') that takes advantage of and enhances the multiple vault password support. If a vault password script basename ends with the name '-client', consider it a vault password script client. A vault password script 'client' just means that the script will take a '--vault-id' command line arg. The previous vault password script (as invoked by --vault-password-file pointing to an executable) takes no args and returns the password on stdout. But it doesnt know anything about --vault-id or multiple vault passwords. The new 'protocol' of the vault password script takes a cli arg ('--vault-id') so that it can lookup that specific vault-id and return it's password. Since existing vault password scripts don't know the new 'protocol', a way to distinguish password scripts that do understand the protocol was needed. The convention now is to consider password scripts that are named like 'something-client.py' (and executable) to be vault password client scripts. The new client scripts get invoked with the '--vault-id' they were requested for. An example: ansible-playbook --vault-id my_vault_id@contrib/vault/vault-keyring-client.py some_playbook.yml That will cause the 'contrib/vault/vault-keyring-client.py' script to be invoked as: contrib/vault/vault-keyring-client.py --vault-id my_vault_id The previous vault-keyring.py password script was extended to become vault-keyring-client.py. It uses the python 'keyring' module to request secrets from various backends. The plain 'vault-keyring.py' script would determine which key id and keyring name to use based on values that had to be set in ansible.cfg. So it was also limited to one keyring name. The new vault-keyring-client.py will request the secret for the vault id provided via the '--vault-id' option. The script can be used without config and can be used for multiple keyring ids (and keyrings). On success, a vault password client script will print the password to stdout and exit with a return code of 0. If the 'client' script can't find a secret for the --vault-id, the script will exit with return code of 2 and print an error to stderr.
2017-10-13 19:23:08 +00:00
def script_is_client(filename):
'''Determine if a vault secret script is a client script that can be given --vault-id args'''
# if password script is 'something-client' or 'something-client.[sh|py|rb|etc]'
# script_name can still have '.' or could be entire filename if there is no ext
script_name, dummy = os.path.splitext(filename)
# TODO: for now, this is entirely based on filename
if script_name.endswith('-client'):
return True
return False
def get_file_vault_secret(filename=None, vault_id=None, encoding=None, loader=None):
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
this_path = os.path.realpath(os.path.expanduser(filename))
if not os.path.exists(this_path):
raise AnsibleError("The vault password file %s was not found" % this_path)
if loader.is_executable(this_path):
Vault secrets script client inc new 'keyring' client (#27669) This adds a new type of vault-password script (a 'client') that takes advantage of and enhances the multiple vault password support. If a vault password script basename ends with the name '-client', consider it a vault password script client. A vault password script 'client' just means that the script will take a '--vault-id' command line arg. The previous vault password script (as invoked by --vault-password-file pointing to an executable) takes no args and returns the password on stdout. But it doesnt know anything about --vault-id or multiple vault passwords. The new 'protocol' of the vault password script takes a cli arg ('--vault-id') so that it can lookup that specific vault-id and return it's password. Since existing vault password scripts don't know the new 'protocol', a way to distinguish password scripts that do understand the protocol was needed. The convention now is to consider password scripts that are named like 'something-client.py' (and executable) to be vault password client scripts. The new client scripts get invoked with the '--vault-id' they were requested for. An example: ansible-playbook --vault-id my_vault_id@contrib/vault/vault-keyring-client.py some_playbook.yml That will cause the 'contrib/vault/vault-keyring-client.py' script to be invoked as: contrib/vault/vault-keyring-client.py --vault-id my_vault_id The previous vault-keyring.py password script was extended to become vault-keyring-client.py. It uses the python 'keyring' module to request secrets from various backends. The plain 'vault-keyring.py' script would determine which key id and keyring name to use based on values that had to be set in ansible.cfg. So it was also limited to one keyring name. The new vault-keyring-client.py will request the secret for the vault id provided via the '--vault-id' option. The script can be used without config and can be used for multiple keyring ids (and keyrings). On success, a vault password client script will print the password to stdout and exit with a return code of 0. If the 'client' script can't find a secret for the --vault-id, the script will exit with return code of 2 and print an error to stderr.
2017-10-13 19:23:08 +00:00
if script_is_client(filename):
display.vvvv('The vault password file %s is a client script.' % filename)
# TODO: pass vault_id_name to script via cli
return ClientScriptVaultSecret(filename=this_path, vault_id=vault_id,
encoding=encoding, loader=loader)
# just a plain vault password script. No args, returns a byte array
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
return ScriptVaultSecret(filename=this_path, encoding=encoding, loader=loader)
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
return FileVaultSecret(filename=this_path, encoding=encoding, loader=loader)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# TODO: mv these classes to a separate file so we don't pollute vault with 'subprocess' etc
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
class FileVaultSecret(VaultSecret):
def __init__(self, filename=None, encoding=None, loader=None):
super(FileVaultSecret, self).__init__()
self.filename = filename
self.loader = loader
self.encoding = encoding or 'utf8'
# We could load from file here, but that is eventually a pain to test
self._bytes = None
self._text = None
@property
def bytes(self):
if self._bytes:
return self._bytes
if self._text:
return self._text.encode(self.encoding)
return None
def load(self):
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
self._bytes = self._read_file(self.filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
def _read_file(self, filename):
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
"""
Read a vault password from a file or if executable, execute the script and
retrieve password from STDOUT
"""
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
# TODO: replace with use of self.loader
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
try:
f = open(filename, "rb")
vault_pass = f.read().strip()
f.close()
except (OSError, IOError) as e:
raise AnsibleError("Could not read vault password file %s: %s" % (filename, e))
b_vault_data, dummy = self.loader._decrypt_if_vault_data(vault_pass, filename)
vault_pass = b_vault_data.strip(b'\r\n')
verify_secret_is_not_empty(vault_pass,
msg='Invalid vault password was provided from file (%s)' % filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
return vault_pass
def __repr__(self):
if self.filename:
return "%s(filename='%s')" % (self.__class__.__name__, self.filename)
return "%s()" % (self.__class__.__name__)
class ScriptVaultSecret(FileVaultSecret):
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
def _read_file(self, filename):
if not self.loader.is_executable(filename):
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
raise AnsibleVaultError("The vault password script %s was not executable" % filename)
Vault secrets script client inc new 'keyring' client (#27669) This adds a new type of vault-password script (a 'client') that takes advantage of and enhances the multiple vault password support. If a vault password script basename ends with the name '-client', consider it a vault password script client. A vault password script 'client' just means that the script will take a '--vault-id' command line arg. The previous vault password script (as invoked by --vault-password-file pointing to an executable) takes no args and returns the password on stdout. But it doesnt know anything about --vault-id or multiple vault passwords. The new 'protocol' of the vault password script takes a cli arg ('--vault-id') so that it can lookup that specific vault-id and return it's password. Since existing vault password scripts don't know the new 'protocol', a way to distinguish password scripts that do understand the protocol was needed. The convention now is to consider password scripts that are named like 'something-client.py' (and executable) to be vault password client scripts. The new client scripts get invoked with the '--vault-id' they were requested for. An example: ansible-playbook --vault-id my_vault_id@contrib/vault/vault-keyring-client.py some_playbook.yml That will cause the 'contrib/vault/vault-keyring-client.py' script to be invoked as: contrib/vault/vault-keyring-client.py --vault-id my_vault_id The previous vault-keyring.py password script was extended to become vault-keyring-client.py. It uses the python 'keyring' module to request secrets from various backends. The plain 'vault-keyring.py' script would determine which key id and keyring name to use based on values that had to be set in ansible.cfg. So it was also limited to one keyring name. The new vault-keyring-client.py will request the secret for the vault id provided via the '--vault-id' option. The script can be used without config and can be used for multiple keyring ids (and keyrings). On success, a vault password client script will print the password to stdout and exit with a return code of 0. If the 'client' script can't find a secret for the --vault-id, the script will exit with return code of 2 and print an error to stderr.
2017-10-13 19:23:08 +00:00
command = self._build_command()
stdout, stderr, p = self._run(command)
self._check_results(stdout, stderr, p)
vault_pass = stdout.strip(b'\r\n')
empty_password_msg = 'Invalid vault password was provided from script (%s)' % filename
verify_secret_is_not_empty(vault_pass,
msg=empty_password_msg)
return vault_pass
def _run(self, command):
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
try:
# STDERR not captured to make it easier for users to prompt for input in their scripts
Vault secrets script client inc new 'keyring' client (#27669) This adds a new type of vault-password script (a 'client') that takes advantage of and enhances the multiple vault password support. If a vault password script basename ends with the name '-client', consider it a vault password script client. A vault password script 'client' just means that the script will take a '--vault-id' command line arg. The previous vault password script (as invoked by --vault-password-file pointing to an executable) takes no args and returns the password on stdout. But it doesnt know anything about --vault-id or multiple vault passwords. The new 'protocol' of the vault password script takes a cli arg ('--vault-id') so that it can lookup that specific vault-id and return it's password. Since existing vault password scripts don't know the new 'protocol', a way to distinguish password scripts that do understand the protocol was needed. The convention now is to consider password scripts that are named like 'something-client.py' (and executable) to be vault password client scripts. The new client scripts get invoked with the '--vault-id' they were requested for. An example: ansible-playbook --vault-id my_vault_id@contrib/vault/vault-keyring-client.py some_playbook.yml That will cause the 'contrib/vault/vault-keyring-client.py' script to be invoked as: contrib/vault/vault-keyring-client.py --vault-id my_vault_id The previous vault-keyring.py password script was extended to become vault-keyring-client.py. It uses the python 'keyring' module to request secrets from various backends. The plain 'vault-keyring.py' script would determine which key id and keyring name to use based on values that had to be set in ansible.cfg. So it was also limited to one keyring name. The new vault-keyring-client.py will request the secret for the vault id provided via the '--vault-id' option. The script can be used without config and can be used for multiple keyring ids (and keyrings). On success, a vault password client script will print the password to stdout and exit with a return code of 0. If the 'client' script can't find a secret for the --vault-id, the script will exit with return code of 2 and print an error to stderr.
2017-10-13 19:23:08 +00:00
p = subprocess.Popen(command, stdout=subprocess.PIPE)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
except OSError as e:
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
msg_format = "Problem running vault password script %s (%s)." \
" If this is not a script, remove the executable bit from the file."
Vault secrets script client inc new 'keyring' client (#27669) This adds a new type of vault-password script (a 'client') that takes advantage of and enhances the multiple vault password support. If a vault password script basename ends with the name '-client', consider it a vault password script client. A vault password script 'client' just means that the script will take a '--vault-id' command line arg. The previous vault password script (as invoked by --vault-password-file pointing to an executable) takes no args and returns the password on stdout. But it doesnt know anything about --vault-id or multiple vault passwords. The new 'protocol' of the vault password script takes a cli arg ('--vault-id') so that it can lookup that specific vault-id and return it's password. Since existing vault password scripts don't know the new 'protocol', a way to distinguish password scripts that do understand the protocol was needed. The convention now is to consider password scripts that are named like 'something-client.py' (and executable) to be vault password client scripts. The new client scripts get invoked with the '--vault-id' they were requested for. An example: ansible-playbook --vault-id my_vault_id@contrib/vault/vault-keyring-client.py some_playbook.yml That will cause the 'contrib/vault/vault-keyring-client.py' script to be invoked as: contrib/vault/vault-keyring-client.py --vault-id my_vault_id The previous vault-keyring.py password script was extended to become vault-keyring-client.py. It uses the python 'keyring' module to request secrets from various backends. The plain 'vault-keyring.py' script would determine which key id and keyring name to use based on values that had to be set in ansible.cfg. So it was also limited to one keyring name. The new vault-keyring-client.py will request the secret for the vault id provided via the '--vault-id' option. The script can be used without config and can be used for multiple keyring ids (and keyrings). On success, a vault password client script will print the password to stdout and exit with a return code of 0. If the 'client' script can't find a secret for the --vault-id, the script will exit with return code of 2 and print an error to stderr.
2017-10-13 19:23:08 +00:00
msg = msg_format % (self.filename, e)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
raise AnsibleError(msg)
stdout, stderr = p.communicate()
Vault secrets script client inc new 'keyring' client (#27669) This adds a new type of vault-password script (a 'client') that takes advantage of and enhances the multiple vault password support. If a vault password script basename ends with the name '-client', consider it a vault password script client. A vault password script 'client' just means that the script will take a '--vault-id' command line arg. The previous vault password script (as invoked by --vault-password-file pointing to an executable) takes no args and returns the password on stdout. But it doesnt know anything about --vault-id or multiple vault passwords. The new 'protocol' of the vault password script takes a cli arg ('--vault-id') so that it can lookup that specific vault-id and return it's password. Since existing vault password scripts don't know the new 'protocol', a way to distinguish password scripts that do understand the protocol was needed. The convention now is to consider password scripts that are named like 'something-client.py' (and executable) to be vault password client scripts. The new client scripts get invoked with the '--vault-id' they were requested for. An example: ansible-playbook --vault-id my_vault_id@contrib/vault/vault-keyring-client.py some_playbook.yml That will cause the 'contrib/vault/vault-keyring-client.py' script to be invoked as: contrib/vault/vault-keyring-client.py --vault-id my_vault_id The previous vault-keyring.py password script was extended to become vault-keyring-client.py. It uses the python 'keyring' module to request secrets from various backends. The plain 'vault-keyring.py' script would determine which key id and keyring name to use based on values that had to be set in ansible.cfg. So it was also limited to one keyring name. The new vault-keyring-client.py will request the secret for the vault id provided via the '--vault-id' option. The script can be used without config and can be used for multiple keyring ids (and keyrings). On success, a vault password client script will print the password to stdout and exit with a return code of 0. If the 'client' script can't find a secret for the --vault-id, the script will exit with return code of 2 and print an error to stderr.
2017-10-13 19:23:08 +00:00
return stdout, stderr, p
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
Vault secrets script client inc new 'keyring' client (#27669) This adds a new type of vault-password script (a 'client') that takes advantage of and enhances the multiple vault password support. If a vault password script basename ends with the name '-client', consider it a vault password script client. A vault password script 'client' just means that the script will take a '--vault-id' command line arg. The previous vault password script (as invoked by --vault-password-file pointing to an executable) takes no args and returns the password on stdout. But it doesnt know anything about --vault-id or multiple vault passwords. The new 'protocol' of the vault password script takes a cli arg ('--vault-id') so that it can lookup that specific vault-id and return it's password. Since existing vault password scripts don't know the new 'protocol', a way to distinguish password scripts that do understand the protocol was needed. The convention now is to consider password scripts that are named like 'something-client.py' (and executable) to be vault password client scripts. The new client scripts get invoked with the '--vault-id' they were requested for. An example: ansible-playbook --vault-id my_vault_id@contrib/vault/vault-keyring-client.py some_playbook.yml That will cause the 'contrib/vault/vault-keyring-client.py' script to be invoked as: contrib/vault/vault-keyring-client.py --vault-id my_vault_id The previous vault-keyring.py password script was extended to become vault-keyring-client.py. It uses the python 'keyring' module to request secrets from various backends. The plain 'vault-keyring.py' script would determine which key id and keyring name to use based on values that had to be set in ansible.cfg. So it was also limited to one keyring name. The new vault-keyring-client.py will request the secret for the vault id provided via the '--vault-id' option. The script can be used without config and can be used for multiple keyring ids (and keyrings). On success, a vault password client script will print the password to stdout and exit with a return code of 0. If the 'client' script can't find a secret for the --vault-id, the script will exit with return code of 2 and print an error to stderr.
2017-10-13 19:23:08 +00:00
def _check_results(self, stdout, stderr, popen):
if popen.returncode != 0:
raise AnsibleError("Vault password script %s returned non-zero (%s): %s" %
(self.filename, popen.returncode, stderr))
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
Vault secrets script client inc new 'keyring' client (#27669) This adds a new type of vault-password script (a 'client') that takes advantage of and enhances the multiple vault password support. If a vault password script basename ends with the name '-client', consider it a vault password script client. A vault password script 'client' just means that the script will take a '--vault-id' command line arg. The previous vault password script (as invoked by --vault-password-file pointing to an executable) takes no args and returns the password on stdout. But it doesnt know anything about --vault-id or multiple vault passwords. The new 'protocol' of the vault password script takes a cli arg ('--vault-id') so that it can lookup that specific vault-id and return it's password. Since existing vault password scripts don't know the new 'protocol', a way to distinguish password scripts that do understand the protocol was needed. The convention now is to consider password scripts that are named like 'something-client.py' (and executable) to be vault password client scripts. The new client scripts get invoked with the '--vault-id' they were requested for. An example: ansible-playbook --vault-id my_vault_id@contrib/vault/vault-keyring-client.py some_playbook.yml That will cause the 'contrib/vault/vault-keyring-client.py' script to be invoked as: contrib/vault/vault-keyring-client.py --vault-id my_vault_id The previous vault-keyring.py password script was extended to become vault-keyring-client.py. It uses the python 'keyring' module to request secrets from various backends. The plain 'vault-keyring.py' script would determine which key id and keyring name to use based on values that had to be set in ansible.cfg. So it was also limited to one keyring name. The new vault-keyring-client.py will request the secret for the vault id provided via the '--vault-id' option. The script can be used without config and can be used for multiple keyring ids (and keyrings). On success, a vault password client script will print the password to stdout and exit with a return code of 0. If the 'client' script can't find a secret for the --vault-id, the script will exit with return code of 2 and print an error to stderr.
2017-10-13 19:23:08 +00:00
def _build_command(self):
return [self.filename]
class ClientScriptVaultSecret(ScriptVaultSecret):
VAULT_ID_UNKNOWN_RC = 2
def __init__(self, filename=None, encoding=None, loader=None, vault_id=None):
super(ClientScriptVaultSecret, self).__init__(filename=filename,
encoding=encoding,
loader=loader)
self._vault_id = vault_id
display.vvvv('Executing vault password client script: %s --vault-id %s' % (filename, vault_id))
Vault secrets script client inc new 'keyring' client (#27669) This adds a new type of vault-password script (a 'client') that takes advantage of and enhances the multiple vault password support. If a vault password script basename ends with the name '-client', consider it a vault password script client. A vault password script 'client' just means that the script will take a '--vault-id' command line arg. The previous vault password script (as invoked by --vault-password-file pointing to an executable) takes no args and returns the password on stdout. But it doesnt know anything about --vault-id or multiple vault passwords. The new 'protocol' of the vault password script takes a cli arg ('--vault-id') so that it can lookup that specific vault-id and return it's password. Since existing vault password scripts don't know the new 'protocol', a way to distinguish password scripts that do understand the protocol was needed. The convention now is to consider password scripts that are named like 'something-client.py' (and executable) to be vault password client scripts. The new client scripts get invoked with the '--vault-id' they were requested for. An example: ansible-playbook --vault-id my_vault_id@contrib/vault/vault-keyring-client.py some_playbook.yml That will cause the 'contrib/vault/vault-keyring-client.py' script to be invoked as: contrib/vault/vault-keyring-client.py --vault-id my_vault_id The previous vault-keyring.py password script was extended to become vault-keyring-client.py. It uses the python 'keyring' module to request secrets from various backends. The plain 'vault-keyring.py' script would determine which key id and keyring name to use based on values that had to be set in ansible.cfg. So it was also limited to one keyring name. The new vault-keyring-client.py will request the secret for the vault id provided via the '--vault-id' option. The script can be used without config and can be used for multiple keyring ids (and keyrings). On success, a vault password client script will print the password to stdout and exit with a return code of 0. If the 'client' script can't find a secret for the --vault-id, the script will exit with return code of 2 and print an error to stderr.
2017-10-13 19:23:08 +00:00
def _run(self, command):
try:
p = subprocess.Popen(command,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE)
except OSError as e:
msg_format = "Problem running vault password client script %s (%s)." \
" If this is not a script, remove the executable bit from the file."
msg = msg_format % (self.filename, e)
raise AnsibleError(msg)
stdout, stderr = p.communicate()
return stdout, stderr, p
def _check_results(self, stdout, stderr, popen):
if popen.returncode == self.VAULT_ID_UNKNOWN_RC:
raise AnsibleError('Vault password client script %s did not find a secret for vault-id=%s: %s' %
(self.filename, self._vault_id, stderr))
if popen.returncode != 0:
raise AnsibleError("Vault password client script %s returned non-zero (%s) when getting secret for vault-id=%s: %s" %
(self.filename, popen.returncode, self._vault_id, stderr))
def _build_command(self):
command = [self.filename]
if self._vault_id:
command.extend(['--vault-id', self._vault_id])
return command
def __repr__(self):
if self.filename:
return "%s(filename='%s', vault_id='%s')" % \
(self.__class__.__name__, self.filename, self._vault_id)
return "%s()" % (self.__class__.__name__)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def match_secrets(secrets, target_vault_ids):
'''Find all VaultSecret objects that are mapped to any of the target_vault_ids in secrets'''
if not secrets:
return []
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
matches = [(vault_id, secret) for vault_id, secret in secrets if vault_id in target_vault_ids]
return matches
def match_best_secret(secrets, target_vault_ids):
'''Find the best secret from secrets that matches target_vault_ids
Since secrets should be ordered so the early secrets are 'better' than later ones, this
just finds all the matches, then returns the first secret'''
matches = match_secrets(secrets, target_vault_ids)
if matches:
return matches[0]
# raise exception?
return None
def match_encrypt_vault_id_secret(secrets, encrypt_vault_id=None):
# See if the --encrypt-vault-id matches a vault-id
display.vvvv('encrypt_vault_id=%s' % encrypt_vault_id)
if encrypt_vault_id is None:
raise AnsibleError('match_encrypt_vault_id_secret requires a non None encrypt_vault_id')
encrypt_vault_id_matchers = [encrypt_vault_id]
encrypt_secret = match_best_secret(secrets, encrypt_vault_id_matchers)
# return the best match for --encrypt-vault-id
if encrypt_secret:
return encrypt_secret
# If we specified a encrypt_vault_id and we couldn't find it, dont
# fallback to using the first/best secret
raise AnsibleVaultError('Did not find a match for --encrypt-vault-id=%s in the known vault-ids %s' % (encrypt_vault_id,
[_v for _v, _vs in secrets]))
def match_encrypt_secret(secrets, encrypt_vault_id=None):
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
'''Find the best/first/only secret in secrets to use for encrypting'''
display.vvvv('encrypt_vault_id=%s' % encrypt_vault_id)
# See if the --encrypt-vault-id matches a vault-id
if encrypt_vault_id:
return match_encrypt_vault_id_secret(secrets,
encrypt_vault_id=encrypt_vault_id)
# Find the best/first secret from secrets since we didnt specify otherwise
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# ie, consider all of the available secrets as matches
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
_vault_id_matchers = [_vault_id for _vault_id, dummy in secrets]
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
best_secret = match_best_secret(secrets, _vault_id_matchers)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# can be empty list sans any tuple
return best_secret
class VaultLib:
def __init__(self, secrets=None):
self.secrets = secrets or []
self.cipher_name = None
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
self.b_version = b'1.2'
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def encrypt(self, plaintext, secret=None, vault_id=None):
2015-08-24 22:49:55 +00:00
"""Vault encrypt a piece of data.
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
:arg plaintext: a text or byte string to encrypt.
2015-08-24 22:49:55 +00:00
:returns: a utf-8 encoded byte str of encrypted data. The string
contains a header identifying this as vault encrypted data and
formatted to newline terminated lines of 80 characters. This is
suitable for dumping as is to a vault file.
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
If the string passed in is a text string, it will be encoded to UTF-8
before encryption.
2015-08-24 22:49:55 +00:00
"""
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
if secret is None:
if self.secrets:
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
dummy, secret = match_encrypt_secret(self.secrets)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
else:
raise AnsibleVaultError("A vault password must be specified to encrypt data")
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
b_plaintext = to_bytes(plaintext, errors='surrogate_or_strict')
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
if is_encrypted(b_plaintext):
raise AnsibleError("input is already encrypted")
if not self.cipher_name or self.cipher_name not in CIPHER_WRITE_WHITELIST:
2015-08-24 22:49:55 +00:00
self.cipher_name = u"AES256"
try:
this_cipher = CIPHER_MAPPING[self.cipher_name]()
except KeyError:
2015-08-24 22:49:55 +00:00
raise AnsibleError(u"{0} cipher could not be found".format(self.cipher_name))
2015-08-24 22:49:55 +00:00
# encrypt data
if vault_id:
display.vvvvv('Encrypting with vault_id "%s" and vault secret %s' % (vault_id, secret))
else:
display.vvvvv('Encrypting without a vault_id using vault secret %s' % secret)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_ciphertext = this_cipher.encrypt(b_plaintext, secret)
2015-08-24 22:49:55 +00:00
# format the data for output to the file
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_vaulttext = format_vaulttext_envelope(b_ciphertext,
self.cipher_name,
vault_id=vault_id)
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
return b_vaulttext
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
def decrypt(self, vaulttext, filename=None):
'''Decrypt a piece of vault encrypted data.
:arg vaulttext: a string to decrypt. Since vault encrypted data is an
ascii text format this can be either a byte str or unicode string.
:kwarg filename: a filename that the data came from. This is only
used to make better error messages in case the data cannot be
decrypted.
:returns: a byte string containing the decrypted data and the vault-id that was used
'''
plaintext, vault_id, vault_secret = self.decrypt_and_get_vault_id(vaulttext, filename=filename)
return plaintext
def decrypt_and_get_vault_id(self, vaulttext, filename=None):
2015-08-24 22:49:55 +00:00
"""Decrypt a piece of vault encrypted data.
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
:arg vaulttext: a string to decrypt. Since vault encrypted data is an
2015-08-24 22:49:55 +00:00
ascii text format this can be either a byte str or unicode string.
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
:kwarg filename: a filename that the data came from. This is only
used to make better error messages in case the data cannot be
decrypted.
:returns: a byte string containing the decrypted data and the vault-id vault-secret that was used
2015-08-24 22:49:55 +00:00
"""
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
b_vaulttext = to_bytes(vaulttext, errors='strict', encoding='utf-8')
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
if self.secrets is None:
raise AnsibleVaultError("A vault password must be specified to decrypt data")
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
if not is_encrypted(b_vaulttext):
msg = "input is not vault encrypted data"
if filename:
msg += "%s is not a vault encrypted file" % to_native(filename)
raise AnsibleError(msg)
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
b_vaulttext, dummy, cipher_name, vault_id = parse_vaulttext_envelope(b_vaulttext,
filename=filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# create the cipher object, note that the cipher used for decrypt can
# be different than the cipher used for encrypt
if cipher_name in CIPHER_WHITELIST:
this_cipher = CIPHER_MAPPING[cipher_name]()
else:
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
raise AnsibleError("{0} cipher could not be found".format(cipher_name))
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_plaintext = None
2015-08-24 22:49:55 +00:00
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
if not self.secrets:
raise AnsibleVaultError('Attempting to decrypt but no vault secrets found')
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# WARNING: Currently, the vault id is not required to match the vault id in the vault blob to
# decrypt a vault properly. The vault id in the vault blob is not part of the encrypted
# or signed vault payload. There is no cryptographic checking/verification/validation of the
# vault blobs vault id. It can be tampered with and changed. The vault id is just a nick
# name to use to pick the best secret and provide some ux/ui info.
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# iterate over all the applicable secrets (all of them by default) until one works...
# if we specify a vault_id, only the corresponding vault secret is checked and
# we check it first.
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
vault_id_matchers = []
vault_id_used = None
vault_secret_used = None
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
if vault_id:
display.vvvvv('Found a vault_id (%s) in the vaulttext' % (vault_id))
vault_id_matchers.append(vault_id)
_matches = match_secrets(self.secrets, vault_id_matchers)
if _matches:
display.vvvvv('We have a secret associated with vault id (%s), will try to use to decrypt %s' % (vault_id, to_text(filename)))
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
else:
display.vvvvv('Found a vault_id (%s) in the vault text, but we do not have a associated secret (--vault-id)' % (vault_id))
# Not adding the other secrets to vault_secret_ids enforces a match between the vault_id from the vault_text and
# the known vault secrets.
if not C.DEFAULT_VAULT_ID_MATCH:
# Add all of the known vault_ids as candidates for decrypting a vault.
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
vault_id_matchers.extend([_vault_id for _vault_id, _dummy in self.secrets if _vault_id != vault_id])
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
matched_secrets = match_secrets(self.secrets, vault_id_matchers)
# for vault_secret_id in vault_secret_ids:
for vault_secret_id, vault_secret in matched_secrets:
display.vvvvv('Trying to use vault secret=(%s) id=%s to decrypt %s' % (vault_secret, vault_secret_id, to_text(filename)))
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
try:
# secret = self.secrets[vault_secret_id]
display.vvvv('Trying secret %s for vault_id=%s' % (vault_secret, vault_secret_id))
b_plaintext = this_cipher.decrypt(b_vaulttext, vault_secret)
if b_plaintext is not None:
vault_id_used = vault_secret_id
vault_secret_used = vault_secret
file_slug = ''
if filename:
file_slug = ' of "%s"' % filename
display.vvvvv('Decrypt%s successful with secret=%s and vault_id=%s' % (file_slug, vault_secret, vault_secret_id))
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
break
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
except AnsibleVaultFormatError as exc:
msg = "There was a vault format error"
if filename:
msg += ' in %s' % (to_text(filename))
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
msg += ': %s' % exc
display.warning(msg)
raise
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
except AnsibleError as e:
display.vvvv('Tried to use the vault secret (%s) to decrypt (%s) but it failed. Error: %s' %
(vault_secret_id, to_text(filename), e))
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
continue
else:
msg = "Decryption failed (no vault secrets were found that could decrypt)"
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
if filename:
msg += " on %s" % to_native(filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
raise AnsibleVaultError(msg)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
if b_plaintext is None:
msg = "Decryption failed"
if filename:
msg += " on %s" % to_native(filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
raise AnsibleError(msg)
return b_plaintext, vault_id_used, vault_secret_used
2015-08-24 22:49:55 +00:00
class VaultEditor:
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def __init__(self, vault=None):
# TODO: it may be more useful to just make VaultSecrets and index of VaultLib objects...
self.vault = vault or VaultLib()
# TODO: mv shred file stuff to it's own class
def _shred_file_custom(self, tmp_path):
""""Destroy a file, when shred (core-utils) is not available
Unix `shred' destroys files "so that they can be recovered only with great difficulty with
specialised hardware, if at all". It is based on the method from the paper
"Secure Deletion of Data from Magnetic and Solid-State Memory",
Proceedings of the Sixth USENIX Security Symposium (San Jose, California, July 22-25, 1996).
We do not go to that length to re-implement shred in Python; instead, overwriting with a block
of random data should suffice.
See https://github.com/ansible/ansible/pull/13700 .
"""
file_len = os.path.getsize(tmp_path)
if file_len > 0: # avoid work when file was empty
max_chunk_len = min(1024 * 1024 * 2, file_len)
passes = 3
with open(tmp_path, "wb") as fh:
for _ in range(passes):
fh.seek(0, 0)
# get a random chunk of data, each pass with other length
chunk_len = random.randint(max_chunk_len // 2, max_chunk_len)
data = os.urandom(chunk_len)
for _ in range(0, file_len // chunk_len):
fh.write(data)
fh.write(data[:file_len % chunk_len])
# FIXME remove this assert once we have unittests to check its accuracy
if fh.tell() != file_len:
raise AnsibleAssertionError()
os.fsync(fh)
def _shred_file(self, tmp_path):
"""Securely destroy a decrypted file
Note standard limitations of GNU shred apply (For flash, overwriting would have no effect
due to wear leveling; for other storage systems, the async kernel->filesystem->disk calls never
guarantee data hits the disk; etc). Furthermore, if your tmp dirs is on tmpfs (ramdisks),
it is a non-issue.
Nevertheless, some form of overwriting the data (instead of just removing the fs index entry) is
a good idea. If shred is not available (e.g. on windows, or no core-utils installed), fall back on
a custom shredding method.
"""
if not os.path.isfile(tmp_path):
# file is already gone
return
try:
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
r = subprocess.call(['shred', tmp_path])
except (OSError, ValueError):
2016-12-11 02:50:09 +00:00
# shred is not available on this system, or some other error occurred.
# ValueError caught because macOS El Capitan is raising an
# exception big enough to hit a limit in python2-2.7.11 and below.
# Symptom is ValueError: insecure pickle when shred is not
# installed there.
r = 1
if r != 0:
# we could not successfully execute unix shred; therefore, do custom shred.
self._shred_file_custom(tmp_path)
os.remove(tmp_path)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def _edit_file_helper(self, filename, secret,
existing_data=None, force_save=False, vault_id=None):
# Create a tempfile
root, ext = os.path.splitext(os.path.realpath(filename))
fd, tmp_path = tempfile.mkstemp(suffix=ext)
os.close(fd)
cmd = self._editor_shell_command(tmp_path)
try:
if existing_data:
self.write_data(existing_data, tmp_path, shred=False)
# drop the user into an editor on the tmp file
subprocess.call(cmd)
except Exception as e:
# whatever happens, destroy the decrypted file
self._shred_file(tmp_path)
raise AnsibleError('Unable to execute the command "%s": %s' % (' '.join(cmd), to_native(e)))
b_tmpdata = self.read_data(tmp_path)
# Do nothing if the content has not changed
if existing_data == b_tmpdata and not force_save:
self._shred_file(tmp_path)
return
# encrypt new data and write out to tmp
# An existing vaultfile will always be UTF-8,
# so decode to unicode here
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_ciphertext = self.vault.encrypt(b_tmpdata, secret, vault_id=vault_id)
self.write_data(b_ciphertext, tmp_path)
# shuffle tmp file into place
self.shuffle_files(tmp_path, filename)
display.vvvvv('Saved edited file "%s" encrypted using %s and vault id "%s"' % (filename, secret, vault_id))
def _real_path(self, filename):
# '-' is special to VaultEditor, dont expand it.
if filename == '-':
return filename
real_path = os.path.realpath(filename)
return real_path
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def encrypt_bytes(self, b_plaintext, secret, vault_id=None):
Vault encrypt string cli (#21024) * Add a vault 'encrypt_string' command. The command will encrypt the string on the command line and print out the yaml block that can be included in a playbook. To be prompted for a string to encrypt: ansible-vault encrypt_string --prompt To specify a string on the command line: ansible-vault encrypt_string "some string to encrypt" To read a string from stdin to encrypt: echo "the plaintext to encrypt" | ansible-vault encrypt_string If a --name or --stdin-name is provided, the output will include that name in yaml key value format: $ ansible-vault encrypt_string "42" --name "the_answer" the_answer: !vault-encrypted | $ANSIBLE_VAULT;1.1;AES256 <vault cipher text here> plaintext provided via prompt, cli, and/or stdin can be mixed: $ ansible-vault encrypt_string "42" --name "the_answer" --prompt Vault password: Variable name (enter for no name): some_variable String to encrypt: microfiber # The encrypted version of variable ("some_variable", the string #1 from the interactive prompt). some_variable: !vault-encrypted | $ANSIBLE_VAULT;1.1;AES256 < vault cipher text here> # The encrypted version of variable ("the_answer", the string #2 from the command line args). the_answer: !vault-encrypted | $ANSIBLE_VAULT;1.1;AES256 < vault cipher text here> Encryption successful * add stdin and prompting to vault 'encrypt_string' * add a --name to encrypt_string to optional specify a var name * prompt for a var name to use with --prompt * add a --stdin-name for the var name for value read from stdin
2017-02-17 15:12:14 +00:00
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_ciphertext = self.vault.encrypt(b_plaintext, secret, vault_id=vault_id)
Vault encrypt string cli (#21024) * Add a vault 'encrypt_string' command. The command will encrypt the string on the command line and print out the yaml block that can be included in a playbook. To be prompted for a string to encrypt: ansible-vault encrypt_string --prompt To specify a string on the command line: ansible-vault encrypt_string "some string to encrypt" To read a string from stdin to encrypt: echo "the plaintext to encrypt" | ansible-vault encrypt_string If a --name or --stdin-name is provided, the output will include that name in yaml key value format: $ ansible-vault encrypt_string "42" --name "the_answer" the_answer: !vault-encrypted | $ANSIBLE_VAULT;1.1;AES256 <vault cipher text here> plaintext provided via prompt, cli, and/or stdin can be mixed: $ ansible-vault encrypt_string "42" --name "the_answer" --prompt Vault password: Variable name (enter for no name): some_variable String to encrypt: microfiber # The encrypted version of variable ("some_variable", the string #1 from the interactive prompt). some_variable: !vault-encrypted | $ANSIBLE_VAULT;1.1;AES256 < vault cipher text here> # The encrypted version of variable ("the_answer", the string #2 from the command line args). the_answer: !vault-encrypted | $ANSIBLE_VAULT;1.1;AES256 < vault cipher text here> Encryption successful * add stdin and prompting to vault 'encrypt_string' * add a --name to encrypt_string to optional specify a var name * prompt for a var name to use with --prompt * add a --stdin-name for the var name for value read from stdin
2017-02-17 15:12:14 +00:00
return b_ciphertext
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def encrypt_file(self, filename, secret, vault_id=None, output_file=None):
# A file to be encrypted into a vaultfile could be any encoding
# so treat the contents as a byte string.
# follow the symlink
filename = self._real_path(filename)
b_plaintext = self.read_data(filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_ciphertext = self.vault.encrypt(b_plaintext, secret, vault_id=vault_id)
self.write_data(b_ciphertext, output_file or filename)
def decrypt_file(self, filename, output_file=None):
# follow the symlink
filename = self._real_path(filename)
ciphertext = self.read_data(filename)
try:
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
plaintext = self.vault.decrypt(ciphertext, filename=filename)
except AnsibleError as e:
raise AnsibleError("%s for %s" % (to_native(e), to_native(filename)))
self.write_data(plaintext, output_file or filename, shred=False)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def create_file(self, filename, secret, vault_id=None):
""" create a new encrypted file """
dirname = os.path.dirname(filename)
if dirname and not os.path.exists(dirname):
display.warning("%s does not exist, creating..." % dirname)
makedirs_safe(dirname)
# FIXME: If we can raise an error here, we can probably just make it
# behave like edit instead.
if os.path.isfile(filename):
raise AnsibleError("%s exists, please use 'edit' instead" % filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
self._edit_file_helper(filename, secret, vault_id=vault_id)
def edit_file(self, filename):
vault_id_used = None
vault_secret_used = None
# follow the symlink
filename = self._real_path(filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_vaulttext = self.read_data(filename)
# vault or yaml files are always utf8
vaulttext = to_text(b_vaulttext)
try:
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# vaulttext gets converted back to bytes, but alas
# TODO: return the vault_id that worked?
plaintext, vault_id_used, vault_secret_used = self.vault.decrypt_and_get_vault_id(vaulttext)
except AnsibleError as e:
raise AnsibleError("%s for %s" % (to_native(e), to_native(filename)))
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# Figure out the vault id from the file, to select the right secret to re-encrypt it
# (duplicates parts of decrypt, but alas...)
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
dummy, dummy, cipher_name, vault_id = parse_vaulttext_envelope(b_vaulttext,
filename=filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# vault id here may not be the vault id actually used for decrypting
# as when the edited file has no vault-id but is decrypted by non-default id in secrets
# (vault_id=default, while a different vault-id decrypted)
# Keep the same vault-id (and version) as in the header
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
if cipher_name not in CIPHER_WRITE_WHITELIST:
2015-08-24 22:49:55 +00:00
# we want to get rid of files encrypted with the AES cipher
self._edit_file_helper(filename, vault_secret_used, existing_data=plaintext,
force_save=True, vault_id=vault_id)
2015-08-24 22:49:55 +00:00
else:
self._edit_file_helper(filename, vault_secret_used, existing_data=plaintext,
force_save=False, vault_id=vault_id)
def plaintext(self, filename):
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_vaulttext = self.read_data(filename)
vaulttext = to_text(b_vaulttext)
try:
plaintext = self.vault.decrypt(vaulttext, filename=filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
return plaintext
except AnsibleError as e:
raise AnsibleVaultError("%s for %s" % (to_native(e), to_native(filename)))
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# FIXME/TODO: make this use VaultSecret
def rekey_file(self, filename, new_vault_secret, new_vault_id=None):
# follow the symlink
filename = self._real_path(filename)
prev = os.stat(filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_vaulttext = self.read_data(filename)
vaulttext = to_text(b_vaulttext)
display.vvvvv('Rekeying file "%s" to with new vault-id "%s" and vault secret %s' %
(filename, new_vault_id, new_vault_secret))
try:
plaintext, vault_id_used, _dummy = self.vault.decrypt_and_get_vault_id(vaulttext)
except AnsibleError as e:
raise AnsibleError("%s for %s" % (to_native(e), to_native(filename)))
# This is more or less an assert, see #18247
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
if new_vault_secret is None:
raise AnsibleError('The value for the new_password to rekey %s with is not valid' % filename)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# FIXME: VaultContext...? could rekey to a different vault_id in the same VaultSecrets
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# Need a new VaultLib because the new vault data can be a different
# vault lib format or cipher (for ex, when we migrate 1.0 style vault data to
# 1.1 style data we change the version and the cipher). This is where a VaultContext might help
# the new vault will only be used for encrypting, so it doesn't need the vault secrets
# (we will pass one in directly to encrypt)
new_vault = VaultLib(secrets={})
b_new_vaulttext = new_vault.encrypt(plaintext, new_vault_secret, vault_id=new_vault_id)
self.write_data(b_new_vaulttext, filename)
2015-11-03 05:27:48 +00:00
# preserve permissions
os.chmod(filename, prev.st_mode)
os.chown(filename, prev.st_uid, prev.st_gid)
display.vvvvv('Rekeyed file "%s" (decrypted with vault id "%s") was encrypted with new vault-id "%s" and vault secret %s' %
(filename, vault_id_used, new_vault_id, new_vault_secret))
def read_data(self, filename):
try:
if filename == '-':
data = sys.stdin.read()
else:
with open(filename, "rb") as fh:
data = fh.read()
except Exception as e:
msg = to_native(e)
if not msg:
msg = repr(e)
raise AnsibleError('Unable to read source file (%s): %s' % (to_native(filename), msg))
return data
# TODO: add docstrings for arg types since this code is picky about that
def write_data(self, data, filename, shred=True):
"""Write the data bytes to given path
This is used to write a byte string to a file or stdout. It is used for
writing the results of vault encryption or decryption. It is used for
saving the ciphertext after encryption and it is also used for saving the
plaintext after decrypting a vault. The type of the 'data' arg should be bytes,
since in the plaintext case, the original contents can be of any text encoding
or arbitrary binary data.
When used to write the result of vault encryption, the val of the 'data' arg
should be a utf-8 encoded byte string and not a text typ and not a text type..
When used to write the result of vault decryption, the val of the 'data' arg
should be a byte string and not a text type.
:arg data: the byte string (bytes) data
:arg filename: filename to save 'data' to.
:arg shred: if shred==True, make sure that the original data is first shredded so that is cannot be recovered.
:returns: None
"""
# FIXME: do we need this now? data_bytes should always be a utf-8 byte string
b_file_data = to_bytes(data, errors='strict')
# get a ref to either sys.stdout.buffer for py3 or plain old sys.stdout for py2
# We need sys.stdout.buffer on py3 so we can write bytes to it since the plaintext
# of the vaulted object could be anything/binary/etc
output = getattr(sys.stdout, 'buffer', sys.stdout)
if filename == '-':
output.write(b_file_data)
else:
if os.path.isfile(filename):
if shred:
self._shred_file(filename)
else:
os.remove(filename)
with open(filename, "wb") as fh:
fh.write(b_file_data)
def shuffle_files(self, src, dest):
prev = None
# overwrite dest with src
if os.path.isfile(dest):
prev = os.stat(dest)
# old file 'dest' was encrypted, no need to _shred_file
os.remove(dest)
shutil.move(src, dest)
# reset permissions if needed
if prev is not None:
# TODO: selinux, ACLs, xattr?
os.chmod(dest, prev.st_mode)
os.chown(dest, prev.st_uid, prev.st_gid)
def _editor_shell_command(self, filename):
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
env_editor = os.environ.get('EDITOR', 'vi')
editor = shlex.split(env_editor)
editor.append(filename)
return editor
########################################
# CIPHERS #
########################################
2015-08-24 22:49:55 +00:00
class VaultAES256:
"""
Vault implementation using AES-CTR with an HMAC-SHA256 authentication code.
Keys are derived using PBKDF2
"""
# http://www.daemonology.net/blog/2009-06-11-cryptographic-right-answers.html
2015-08-24 22:49:55 +00:00
# Note: strings in this class should be byte strings by default.
def __init__(self):
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
if not HAS_CRYPTOGRAPHY and not HAS_PYCRYPTO:
raise AnsibleError(NEED_CRYPTO_LIBRARY)
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
@staticmethod
def _create_key_cryptography(b_password, b_salt, key_length, iv_length):
kdf = PBKDF2HMAC(
algorithm=hashes.SHA256(),
length=2 * key_length + iv_length,
salt=b_salt,
iterations=10000,
backend=CRYPTOGRAPHY_BACKEND)
b_derivedkey = kdf.derive(b_password)
return b_derivedkey
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
@staticmethod
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
def _pbkdf2_prf(p, s):
hash_function = SHA256_pycrypto
return HMAC_pycrypto.new(p, s, hash_function).digest()
@classmethod
def _create_key_pycrypto(cls, b_password, b_salt, key_length, iv_length):
# make two keys and one iv
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
b_derivedkey = PBKDF2_pycrypto(b_password, b_salt, dkLen=(2 * key_length) + iv_length,
count=10000, prf=cls._pbkdf2_prf)
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
return b_derivedkey
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
@classmethod
def _gen_key_initctr(cls, b_password, b_salt):
# 16 for AES 128, 32 for AES256
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
key_length = 32
if HAS_CRYPTOGRAPHY:
# AES is a 128-bit block cipher, so IVs and counter nonces are 16 bytes
iv_length = algorithms.AES.block_size // 8
b_derivedkey = cls._create_key_cryptography(b_password, b_salt, key_length, iv_length)
b_iv = b_derivedkey[(key_length * 2):(key_length * 2) + iv_length]
elif HAS_PYCRYPTO:
# match the size used for counter.new to avoid extra work
iv_length = 16
b_derivedkey = cls._create_key_pycrypto(b_password, b_salt, key_length, iv_length)
b_iv = hexlify(b_derivedkey[(key_length * 2):(key_length * 2) + iv_length])
else:
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
raise AnsibleError(NEED_CRYPTO_LIBRARY + '(Detected in initctr)')
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
b_key1 = b_derivedkey[:key_length]
b_key2 = b_derivedkey[key_length:(key_length * 2)]
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
return b_key1, b_key2, b_iv
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
@staticmethod
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
def _encrypt_cryptography(b_plaintext, b_key1, b_key2, b_iv):
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
cipher = C_Cipher(algorithms.AES(b_key1), modes.CTR(b_iv), CRYPTOGRAPHY_BACKEND)
encryptor = cipher.encryptor()
padder = padding.PKCS7(algorithms.AES.block_size).padder()
b_ciphertext = encryptor.update(padder.update(b_plaintext) + padder.finalize())
b_ciphertext += encryptor.finalize()
# COMBINE SALT, DIGEST AND DATA
hmac = HMAC(b_key2, hashes.SHA256(), CRYPTOGRAPHY_BACKEND)
hmac.update(b_ciphertext)
b_hmac = hmac.finalize()
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
return to_bytes(hexlify(b_hmac), errors='surrogate_or_strict'), hexlify(b_ciphertext)
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
@staticmethod
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
def _encrypt_pycrypto(b_plaintext, b_key1, b_key2, b_iv):
# PKCS#7 PAD DATA http://tools.ietf.org/html/rfc5652#section-6.3
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
bs = AES_pycrypto.block_size
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
padding_length = (bs - len(b_plaintext) % bs) or bs
b_plaintext += to_bytes(padding_length * chr(padding_length), encoding='ascii', errors='strict')
# COUNTER.new PARAMETERS
# 1) nbits (integer) - Length of the counter, in bits.
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
# 2) initial_value (integer) - initial value of the counter. "iv" from _gen_key_initctr
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
ctr = Counter_pycrypto.new(128, initial_value=int(b_iv, 16))
# AES.new PARAMETERS
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
# 1) AES key, must be either 16, 24, or 32 bytes long -- "key" from _gen_key_initctr
# 2) MODE_CTR, is the recommended mode
# 3) counter=<CounterObject>
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
cipher = AES_pycrypto.new(b_key1, AES_pycrypto.MODE_CTR, counter=ctr)
# ENCRYPT PADDED DATA
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
b_ciphertext = cipher.encrypt(b_plaintext)
# COMBINE SALT, DIGEST AND DATA
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
hmac = HMAC_pycrypto.new(b_key2, b_ciphertext, SHA256_pycrypto)
return to_bytes(hmac.hexdigest(), errors='surrogate_or_strict'), hexlify(b_ciphertext)
@classmethod
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def encrypt(cls, b_plaintext, secret):
if secret is None:
raise AnsibleVaultError('The secret passed to encrypt() was None')
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
b_salt = os.urandom(32)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
b_password = secret.bytes
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
b_key1, b_key2, b_iv = cls._gen_key_initctr(b_password, b_salt)
if HAS_CRYPTOGRAPHY:
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
b_hmac, b_ciphertext = cls._encrypt_cryptography(b_plaintext, b_key1, b_key2, b_iv)
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
elif HAS_PYCRYPTO:
pylint fixes for vault related code (#27721) * rm unneeded parens following assert * rm unused parse_vaulttext_envelope from yaml.constructor * No longer need index/enumerate over vault_ids * rm unnecessary else * rm unused VaultCli.secrets * rm unused vault_id arg on VaultAES.decrypt() pylint: Unused argument 'vault_id' pylint: Unused parse_vaulttext_envelope imported from ansible.parsing.vault pylint: Unused variable 'index' pylint: Unnecessary parens after 'assert' keyword pylint: Unnecessary "else" after "return" (no-else-return) pylint: Attribute 'editor' defined outside __init__ * use 'dummy' for unused variables instead of _ Based on pylint unused variable warnings. Existing code use '_' for this, but that is old and busted. The hot new thing is 'dummy'. It is so fetch. Except for where we get warnings for reusing the 'dummy' var name inside of a list comprehension. * Add super().__init__ call to PromptVaultSecret.__init__ pylint: __init__ method from base class 'VaultSecret' is not called (super-init-not-called) * Make FileVaultSecret.read_file reg method again The base class read_file() doesnt need self but the sub classes do. Rm now unneeded loader arg to read_file() * Fix err msg string literal that had no effect pylint: String statement has no effect The indent on the continuation of the msg_format was wrong so the second half was dropped. There was also no need to join() filename (copy/paste from original with a command list I assume...) * Use local cipher_name in VaultEditor.edit_file not instance pylint: Unused variable 'cipher_name' pylint: Unused variable 'b_ciphertext' Use the local cipher_name returned from parse_vaulttext_envelope() instead of the instance self.cipher_name var. Since there is only one valid cipher_name either way, it was equilivent, but it will not be with more valid cipher_names * Rm unused b_salt arg on VaultAES256._encrypt* pylint: Unused argument 'b_salt' Previously the methods computed the keys and iv themselves so needed to be passed in the salt, but now the key/iv are built before and passed in so b_salt arg is not used anymore. * rm redundant import of call from subprocess pylint: Imports from package subprocess are not grouped use via subprocess module now instead of direct import. * self._bytes is set in super init now, rm dup * Make FileVaultSecret.read_file() -> _read_file() _read_file() is details of the implementation of load(), so now 'private'.
2017-08-08 20:10:03 +00:00
b_hmac, b_ciphertext = cls._encrypt_pycrypto(b_plaintext, b_key1, b_key2, b_iv)
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
else:
raise AnsibleError(NEED_CRYPTO_LIBRARY + '(Detected in encrypt)')
b_vaulttext = b'\n'.join([hexlify(b_salt), b_hmac, b_ciphertext])
# Unnecessary but getting rid of it is a backwards incompatible vault
# format change
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
b_vaulttext = hexlify(b_vaulttext)
return b_vaulttext
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
@classmethod
def _decrypt_cryptography(cls, b_ciphertext, b_crypted_hmac, b_key1, b_key2, b_iv):
# b_key1, b_key2, b_iv = self._gen_key_initctr(b_password, b_salt)
# EXIT EARLY IF DIGEST DOESN'T MATCH
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
hmac = HMAC(b_key2, hashes.SHA256(), CRYPTOGRAPHY_BACKEND)
hmac.update(b_ciphertext)
try:
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
hmac.verify(_unhexlify(b_crypted_hmac))
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
except InvalidSignature as e:
raise AnsibleVaultError('HMAC verification failed: %s' % e)
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
cipher = C_Cipher(algorithms.AES(b_key1), modes.CTR(b_iv), CRYPTOGRAPHY_BACKEND)
decryptor = cipher.decryptor()
unpadder = padding.PKCS7(128).unpadder()
b_plaintext = unpadder.update(
decryptor.update(b_ciphertext) + decryptor.finalize()
) + unpadder.finalize()
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
return b_plaintext
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
@staticmethod
def _is_equal(b_a, b_b):
"""
Comparing 2 byte arrrays in constant time
to avoid timing attacks.
It would be nice if there was a library for this but
hey.
"""
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
if not (isinstance(b_a, binary_type) and isinstance(b_b, binary_type)):
raise TypeError('_is_equal can only be used to compare two byte strings')
# http://codahale.com/a-lesson-in-timing-attacks/
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
if len(b_a) != len(b_b):
return False
result = 0
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
for b_x, b_y in zip(b_a, b_b):
2015-06-03 17:24:35 +00:00
if PY3:
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
result |= b_x ^ b_y
2015-06-03 17:24:35 +00:00
else:
Many Cleanups to vault * Make is_encrypted_file handle both files opened in text and binary mode On python3, by default files are opened in text mode. Since we know the encoding of vault files (and especially the header which is the first set of bytes) we can decide whether the file is an encrypted vault file in either case. * Fix is_encrypted_file not resetting the file position * Update is_encrypted_file to check that all the data in the file is ascii * For is_encrypted_file(), add start_pos and count parameters This allows callers to specify reading vaulttext from the middle of a file if necessary. * Combine VaultLib.encrypt() and VaultLib.encrypt_bytestring() * Change vault's is_encrypted() to take either text or byte strings and to return False if any part of the data is non-ascii. * Remove unnecessary use of six.b * Vault Cipher: mark a few methods as private. * VaultAES256._is_equal throws a TypeError if given non byte strings * Make VaultAES256 methods that don't need self staticmethods and classmethods * Mark VaultAES and is_encrypted as deprecated * Get rid of VaultFile (unused and feature implemented in a different way) * Normalize variable and parameter names on plaintext, ciphertext, vaulttext * Normalize variable and parameter names on "b_" prefix when dealing with bytes * Test changes: * Remove redundant tests( both checking the same byte string) * Fix use of format string without format operator * Enable vault editor tests on python3 * Initialize the vault_cipher for VaultAES256 testing in setUp() * Make assertTrue and assertFalse take the actual method calls for better error messages. * Test that non-ascii byte strings compare correctly. * Test that unicode strings and ints raise TypeError * Test-specific: * Removed test_methods_exist(). We only have one VaultLib so the implementation is the assurance that the methods exist. (Can use an abc for this if it changes). * Add tests for both byte string and text string input where the API takes either. * Convert "assert" to unittest assert functions or add a custom message where that will make failures easier to debug. * Move instantiating the VaultLib into setUp().
2016-09-13 14:30:17 +00:00
result |= ord(b_x) ^ ord(b_y)
return result == 0
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
@classmethod
def _decrypt_pycrypto(cls, b_ciphertext, b_crypted_hmac, b_key1, b_key2, b_iv):
# EXIT EARLY IF DIGEST DOESN'T MATCH
hmac_decrypt = HMAC_pycrypto.new(b_key2, b_ciphertext, SHA256_pycrypto)
if not cls._is_equal(b_crypted_hmac, to_bytes(hmac_decrypt.hexdigest())):
return None
# SET THE COUNTER AND THE CIPHER
ctr = Counter_pycrypto.new(128, initial_value=int(b_iv, 16))
cipher = AES_pycrypto.new(b_key1, AES_pycrypto.MODE_CTR, counter=ctr)
# DECRYPT PADDED DATA
b_plaintext = cipher.decrypt(b_ciphertext)
# UNPAD DATA
if PY3:
padding_length = b_plaintext[-1]
else:
padding_length = ord(b_plaintext[-1])
b_plaintext = b_plaintext[:-padding_length]
return b_plaintext
@classmethod
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
def decrypt(cls, b_vaulttext, secret):
Better handling of malformed vault data envelope (#32515) * Better handling of malformed vault data envelope If an embedded vaulted variable ('!vault' in yaml) had an invalid format, it would eventually cause an error for seemingly unrelated reasons. "Invalid" meaning not valid hexlify (extra chars, non-hex chars, etc). For ex, if a host_vars file had invalid vault format variables, on py2, it would cause an error like: 'ansible.vars.hostvars.HostVars object' has no attribute u'broken.example.com' Depending on where the invalid vault is, it could also cause "VARIABLE IS NOT DEFINED!". The behavior can also change if ansible-playbook is py2 or py3. Root cause is errors from binascii.unhexlify() not being handled consistently. Fix is to add a AnsibleVaultFormatError exception and raise it on any unhexlify() errors and to handle it properly elsewhere. Add a _unhexlify() that try/excepts around a binascii.unhexlify() and raises an AnsibleVaultFormatError on invalid vault data. This is so the same exception type is always raised for this case. Previous it was different between py2 and py3. binascii.unhexlify() raises a binascii.Error if the hexlified blobs in a vault data blob are invalid. On py2, binascii.Error is a subclass of Exception. On py3, binascii.Error is a subclass of TypeError When decrypting content of vault encrypted variables, if a binascii.Error is raised it propagates up to playbook.base.Base.post_validate(). post_validate() handles exceptions for TypeErrors but not for base Exception subclasses (like py2 binascii.Error). * Add a display.warning on vault format errors * Unit tests for _unhexlify, parse_vaulttext* * Add intg test cases for invalid vault formats Fixes #28038
2017-11-10 19:24:56 +00:00
b_ciphertext, b_salt, b_crypted_hmac = parse_vaulttext(b_vaulttext)
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# TODO: would be nice if a VaultSecret could be passed directly to _decrypt_*
# (move _gen_key_initctr() to a AES256 VaultSecret or VaultContext impl?)
# though, likely needs to be python cryptography specific impl that basically
# creates a Cipher() with b_key1, a Mode.CTR() with b_iv, and a HMAC() with sign key b_key2
b_password = secret.bytes
Cyptography pr 20566 rebase (#25560) Make pyca/cryptography the preferred backend for cryptographic needs (mainly vault) falling back to pycrypto pyca/cryptography is already implicitly a dependency in many cases through paramiko (2.0+) as well as the new openssl_publickey module, which requires pyOpenSSL 16.0+. Additionally, pyca/cryptography is an optional dep for better performance with vault already. This commit leverages cryptography's padding, constant time comparisons, and CBC/CTR modes to reduce the amount of code ansible needs to maintain. * Handle wrong password given for VaultAES format * Do not display deprecation warning for cryptography on python-2.6 * Namespace all of the pycrypto imports and always import them Makes unittests better and the code less likely to get stupid mistakes (like using HMAC from cryptogrpahy when the one from pycrypto is needed) * Add back in atfork since we need pycrypto to reinitialize its RNG just in case we're being used with old paramiko * contrib/inventory/gce: Remove spurious require on pycrypto (cherry picked from commit 9e16b9db275263b3ea8d1b124966fdebfc9ab271) * Add cryptography to ec2_win_password module requirements * Fix python3 bug which would pass text strings to a function which requires byte strings. * Attempt to add pycrypto version to setup deps * Change hacking README for dual pycrypto/cryptography * update dependencies for various CI scripts * additional CI dockerfile/script updates * add paramiko to the windows and sanity requirement set This is needed because ansible lists it as a requirement. Previously the missing dep wasn't enforced, but cryptography imports pkg_resources so you can't ignore a requirement any more * Add integration test cases for old vault and for wrong passwords * helper script for manual testing of pycrypto/cryptography * Skip the pycrypto tests so that users without it installed can still run the unittests * Run unittests for vault with both cryptography and pycrypto backend
2017-06-27 13:00:15 +00:00
b_key1, b_key2, b_iv = cls._gen_key_initctr(b_password, b_salt)
if HAS_CRYPTOGRAPHY:
b_plaintext = cls._decrypt_cryptography(b_ciphertext, b_crypted_hmac, b_key1, b_key2, b_iv)
elif HAS_PYCRYPTO:
b_plaintext = cls._decrypt_pycrypto(b_ciphertext, b_crypted_hmac, b_key1, b_key2, b_iv)
else:
raise AnsibleError(NEED_CRYPTO_LIBRARY + '(Detected in decrypt)')
return b_plaintext
Support multiple vault passwords (#22756) Fixes #13243 ** Add --vault-id to name/identify multiple vault passwords Use --vault-id to indicate id and path/type --vault-id=prompt # prompt for default vault id password --vault-id=myorg@prompt # prompt for a vault_id named 'myorg' --vault-id=a_password_file # load ./a_password_file for default id --vault-id=myorg@a_password_file # load file for 'myorg' vault id vault_id's are created implicitly for existing --vault-password-file and --ask-vault-pass options. Vault ids are just for UX purposes and bookkeeping. Only the vault payload and the password bytestring is needed to decrypt a vault blob. Replace passing password around everywhere with a VaultSecrets object. If we specify a vault_id, mention that in password prompts Specifying multiple -vault-password-files will now try each until one works ** Rev vault format in a backwards compatible way The 1.2 vault format adds the vault_id to the header line of the vault text. This is backwards compatible with older versions of ansible. Old versions will just ignore it and treat it as the default (and only) vault id. Note: only 2.4+ supports multiple vault passwords, so while earlier ansible versions can read the vault-1.2 format, it does not make them magically support multiple vault passwords. use 1.1 format for 'default' vault_id Vaulted items that need to include a vault_id will be written in 1.2 format. If we set a new DEFAULT_VAULT_IDENTITY, then the default will use version 1.2 vault will only use a vault_id if one is specified. So if none is specified and C.DEFAULT_VAULT_IDENTITY is 'default' we use the old format. ** Changes/refactors needed to implement multiple vault passwords raise exceptions on decrypt fail, check vault id early split out parsing the vault plaintext envelope (with the sha/original plaintext) to _split_plaintext_envelope() some cli fixups for specifying multiple paths in the unfrack_paths optparse callback fix py3 dict.keys() 'dict_keys object is not indexable' error pluralize cli.options.vault_password_file -> vault_password_files pluralize cli.options.new_vault_password_file -> new_vault_password_files pluralize cli.options.vault_id -> cli.options.vault_ids ** Add a config option (vault_id_match) to force vault id matching. With 'vault_id_match=True' and an ansible vault that provides a vault_id, then decryption will require that a matching vault_id is required. (via --vault-id=my_vault_id@password_file, for ex). In other words, if the config option is true, then only the vault secrets with matching vault ids are candidates for decrypting a vault. If option is false (the default), then all of the provided vault secrets will be selected. If a user doesn't want all vault secrets to be tried to decrypt any vault content, they can enable this option. Note: The vault id used for the match is not encrypted or cryptographically signed. It is just a label/id/nickname used for referencing a specific vault secret.
2017-07-28 19:20:58 +00:00
# Keys could be made bytes later if the code that gets the data is more
# naturally byte-oriented
CIPHER_MAPPING = {
u'AES256': VaultAES256,
}