action plugins will now skip _fixup_perms for Powershell. We'll have to come up with another way to do this at some point, but it's not necessary yet since we don't support become on Windows. Also added NotImplementedError throws to chmod/chown/set_facl operations on Powershell (instead of returning '') in case anyone tries to use them in the future.
fixes#15312
* Ziploader proof of concept (jimi-c)
* Cleanups to proof of concept ziploader branch:
* python3 compatible base64 encoding
* zipfile compression (still need to enable toggling this off for
systems without zlib support in python)
* Allow non-wildcard imports (still need to make this recusrsive so that
we can have module_utils code that imports other module_utils code.)
* Better tracebacks: module filename is kept and module_utils directory
is kept so that tracebacks show the real filenames that the errors
appear in.
* Make sure we import modules that are used into the module_utils files that they are used in.
* Set ansible version in a more pythonic way for ziploader than we were doing in module replacer
* Make it possible to set the module compression as an inventory var
This may be necessary on systems where python has been compiled without
zlib compression.
* Refactoring of module_common code:
* module replacer only replaces values that make sense for that type of
file (example: don't attempt to replace python imports if we're in
a powershell module).
* Implement configurable shebang support for ziploader wrapper
* Implement client-side constants (for SELINUX_SPECIAL_FS and SYSLOG)
via environment variable.
* Remove strip_comments param as we're never going to use it (ruins line
numbering)
* Don't repeat ourselves about detecting REPLACER
* Add an easy way to debug
* Port test-module to the ziploader-aware modify_module()
* strip comments and blank lines from the wrapper so we send less over the wire.
* Comments cleanup
* Remember to output write the module line itself in powershell modules
* for line in lines strips the newlines so we have to add them back in
The changes to chown/chmod were broken on Mac (-R was being appended to the end of the command- OSX requires it before the file list).
A number of base action remote setup commands were also blindly proceeding without checking for success. Added error raises for unrecoverable failure cases.
- added new function for action plugins this avoids the very fragile checksum code that is shell dependant.
- ported copy module to it
- converted assemble to new stat function
- some corrections and ported temlpate
- updated old checksum function to use new stat one under the hood
- documented revamped remote checksum method
this was taken out in an effort to default to the user's shell but creates issues as this is not known ahead of time
and its painful to set executable and shell_type for all servers, it should only be needed for those that restrict the user
to specific shells and when /bin/sh is not available. raw and command may still bypass this by explicitly passing None.
fixes#13882
still conditional
This is because we pass arguments to non-newstyle modules via an
external file. If we pipeline, then the interpreter thinks it has to
run the arguments as the script instead of what is piped in via stdin.
* now module errors clearly state msg=MODULE FAILURE
* module's stdout and stderr go into module_stdout and module_stderr keys
which only appear during parsing failure
* invocation module_args are deleted from results provided by action
plugin as errors can keep us from overwriting and then disclosing info that
was meant to be kept hidden due to no_log
* fixed invocation module_args set by basic.py as it was creating different
keys as the invocation in action plugin base.
* results now merge
This was added in 1.9 and 2.0 tried to copy, but since it cannot
obey no_log restrictions I commented it out. I did not remove as
it is still very useful for module invocation debugging.
Environments were not being templated individually, so a variable environment
value was causing the exception regarding dicts to be hit. Also, environments
as inherited were coming through with the tasks listed first, followed by the
parents, so they were being merged backwards. Reversing the list of environments
fixed this.
We were logging the command to be executed many times, which made debug
logs very hard to read. Now we do it only once.
Also makes the logged ssh command line cut-and-paste-able (the lack of
which has confused a number of people by now; the problem being that we
pass the command as a single argument to execve(), so it doesn't need an
extra level of quoting as it does when you try to run it by hand).
Pipelining is a *significant* performance benefit, because each task can
be completed with a single SSH connection (vs. one ssh connection at the
start to mkdir, plus one sftp and one ssh per task).
Pipelining is disabled by default in Ansible because it conflicts with
the use of sudo if 'Defaults requiretty' is set in /etc/sudoers (as it
is on Red Hat) and su (which always requires a tty).
We can (and already do) make sudo/su happy by using "ssh -t" to allocate
a tty, but then the python interpreter goes into interactive mode and is
unhappy with module source being written to its stdin, per the following
comment from connections/ssh.py:
# we can only use tty when we are not pipelining the modules.
# piping data into /usr/bin/python inside a tty automatically
# invokes the python interactive-mode but the modules are not
# compatible with the interactive-mode ("unexpected indent"
# mainly because of empty lines)
Instead of the (current) drastic solution of turning off pipelining when
we use a tty, we can instead use a tty but suppress the behaviour of the
Python interpreter to switch to interactive mode. The easiest way to do
this is to make its stdin *not* be a tty, e.g. with cat|python.
This works, but there's a problem: ssh will ignore -t if its input isn't
really a tty. So we could open a pseudo-tty and use that as ssh's stdin,
but if we then write Python source into it, it's all echoed back to us
(because we're a tty). So we have to use -tt to force tty allocation; in
that case, however, ssh puts the tty into "raw" mode (~ICANON), so there
is no good way for the process on the other end to detect EOF on stdin.
So if we do:
echo -e "print('hello world')\n"|ssh -tt someho.st "cat|python"
…it hangs forever, because cat keeps on reading input even after we've
closed our pipe into ssh's stdin. We can get around this by writing a
special __EOF__ marker after writing in_data, and doing this:
echo -e "print('hello world')\n__EOF__\n"|ssh -tt someho.st "sed -ne '/__EOF__/q' -e p|python"
This works fine, but in fact I use a clever python one-liner by mgedmin
to achieve the same effect without depending on sed (at the expense of a
much longer command line, alas; Python really isn't one-liner-friendly).
We also enable pipelining by default as a consequence.
Matt Davis
Toshio Kuratomi
James Cammarata
Brian Coca
nitzmahone
Toshio Kuratomi
Brian Coca
Matthew Stoltenberg
Matt Martz
Abhijit Menon-Sen
Luca Berruti
Sebastien Couture