When supplying a sudo password to a server that uses passwordless sudo,
we should not throw away useful stdout and stderr. This is particularly
important for modules that perform md5 checks as part of the pre module
execution.
Currently, ansible -vvvv show the following:
hostname EXEC ['ssh', '-C', '-vvv', '-o', 'ControlMaster=auto',
This is not good if someone want to cut and paste it to see if something
is wrong with the command line, so join the dictionnary strings with spaces
to have 1 string in the output.
Addresses multiple issues when using su on freebsd including
* su prompt differs between platforms, so turned that check into a
regex comparison instead of a simple string comparison
* not using '-c' after su causes problems, so added that for all
platforms
* fixed quoting issues due to multiple uses of '-c' introduced by
the above fix
Fixes#7503Fixes#7507
If someone add ssh_args = " " to his .ansible.cfg, it will result into
strange failure later :
<server.example.org> ESTABLISH CONNECTION FOR USER: misc
<server.example.org> REMOTE_MODULE ping
<server.example.org> EXEC ['ssh', '-C', '-tt', '-q', ' ', '-o', 'KbdInteractiveAuthentication=no',
'-o', 'PreferredAuthentications=gssapi-with-mic,gssapi-keyex,hostbased,publickey', '-o', 'PasswordAuthentication=no',
'-o', 'ConnectTimeout=10', 'server.example.org', "/bin/sh -c 'mkdir -p /tmp/ansible-tmp-1397947711.21-5932460998838
&& chmod a+rx /tmp/ansible-tmp-1397947711.21-5932460998838 && echo /tmp/ansible-tmp-1397947711.21-5932460998838'"]
server.example.org | FAILED => SSH encountered an unknown error during the connection. We recommend you re-run the
command using -vvvv, which will enable SSH debugging output to help diagnose the issue
The root cause is the empty string between -q and -o, who kinda break mkdir.
Any other module is able to detect a dark host, but raw was treating 255
as a return code from the module execution, rather from the connection
attempt. This change allows 255 to be treated as a connection failure
when using the raw module.
We break the read while loop after waiting "the end of the process" and
the pipes are empty, otherwise we do another select that waits all the
timeout.
In particular, do not rely on the $USER environment variable always existing.
tmux for example seems to clear it, causing lots of invalid messages:
"previous known host file not found"
This broke in commit 80fd22dc, but instead of reverting that commit, we now
fall back to expanding just ~ when $USER is not set.
Resolves issue #5082
Code as it was would hit a scenario where one of the FDs was not ready for
reading the first time through -- but p.poll() would show the process as
complete. This would cause ansible to continue on, while leaving some content
left in a pipe.
The other scenario -- the one that causes the unclosed quote, is if we go
through select.select() and we do get stdout in the ready for reading -- we
read from it (9000 bytes), but that's not all that is there. Again we'd get to
the p.poll() check and it would be indeed not none, but we would have left some
of stdout on the FD and thus the json blob would be malformed.
Tested with and without full ssh debugging.
Tested with and without ControlPersist
Tested with and without ControlPersist sockets already created
Toshio Kuratomi
Bruno BAILLUET
Matt Martz
Michael Scherer
James Cammarata
Thomas Mangin
Eri Bastos
Marco Re
Michael Scherer
jctanner
BoscoMW
Till Maas
Jesse Keating
Cristian Ciupitu
Iordan Iordanov
Luca Berruti
Heikki Hokkanen
Paul Durivage
James Cammarata
jeromew
Michael DeHaan