741e99f698
ansible 2.10 no longer has a at the end of the error msg... 🤦
4 years ago
95c43ec9fc
fixed issue of switching between mitogen and non-mitogen strategies
4 years ago
e30e84334e
remove synchronize fail test for azure
4 years ago
139b9560bc
print didn't work because verbosity, throw valueerror to see
4 years ago
9bd35adcfb
more debugging, synchronize is being weird on azure
4 years ago
ff8a276186
turn off failing Ansible-only tests for now, also raising errors to see what Azure is gonna do with collections
4 years ago
f757dbcb82
removed duplicate install and added debug dump of collection loading to see what tests are doing
4 years ago
41a13ebce2
hopefully this also fails the same way
4 years ago
5aedb5f157
add missing collections 🤦
4 years ago
9857dfea5c
verify collection is working as expected
4 years ago
a40c28d93f
can't replicate but think it's because synchronize is now a collection
4 years ago
49dd8eee1a
figure out what synchronize is now
4 years ago
c6d42212dd
add some debugging info, was able to run the failed synchronize test locally just fine using test framework, not sure what's going on
4 years ago
f91cbf4d00
test cleanup and trying to replicate synchronize fails
4 years ago
f994807d74
ansible 2.10 renamed ping module to reflect legacy builtin collection
4 years ago
03438271bb
able to remove the hack on ActionBase after all
4 years ago
5215646c8a
code cleanup
4 years ago
0b421e0d3c
able to run docker_container installed via 'ansible-galaxy collection install community.general'
4 years ago
45797a0d34
able to send collections paths to master with very little change to core Mitogen, need to ensure imports work properly now though
4 years ago
6e51f1a184
found a way to load collections without affecting mitogen core code
4 years ago
8d3ee26079
code cleanup
4 years ago
ca94751f15
remove hack
4 years ago
f1bdc39047
added note about breaking backwards compat
4 years ago
1d13df718a
connection_loader.get isn't called anymore, it's connection_loader.get_with_context now
4 years ago
6ba08097b6
more notes, strategy plugin is being called but Mitogen's method overrides aren't being triggered
4 years ago
ca4e8116b7
TODO: turns out ansible 2.10 doesn't run Mitogen like it used to; was running old ansible version before because ansible-base didn't override everything. Did a fresh uninstall of ansible and installed 2.10.0 and Mitogen's connection monkeypatching isn't working
4 years ago
1bd4b8afcd
much thinking needs to be done regarding how to handle not requiring sshpass...
4 years ago
6ac9168d55
need to get around sshpass check here somehow: https://github.com/ansible/ansible/blob/v2.10.0/lib/ansible/plugins/connection/ssh.py#L577
4 years ago
583f540889
added comments
4 years ago
e34cf8667f
Merge branch 'master' into collectionsSupport
4 years ago
81076c9da8
fixes setup module relative import fail on some pythons
5 years ago
376d8d0fab
remove old hacks; ansible_collections is available at time of invoker but not later
5 years ago
955e77c5db
WIP: able to load subdirs but now need to treat them as submodules properly
5 years ago
ab55d05267
all in on ansible 2.10+ for collections support
5 years ago
c92a9ace41
bump max ansible version
5 years ago
ac1e72eec4
Merge branch 'master' into collectionsSupport
5 years ago
e632310fc4
no tmpdir to remove for old ansible versions in fetch command
5 years ago
ddc1eebec8
able to load collection but the mitogen master proc was unable to serve it
5 years ago
e99d63f4fc
able to load collections but they don't do anything because no actionmodulemixin, need to tweak how this works
5 years ago
acde13f9d6
handles a 'wait_for_connection' call right after a task caused a shutdown
5 years ago
5bf327649f
add missing / to python paths
5 years ago
389cee8c70
fix assertion of python_path, it should be always what 'discovered_interpreter' is
5 years ago
c959ce010b
created temp action obj for 'meta: reset_connection' python interpreter discovery
5 years ago
5b8c9da769
_execute_meta doesn't have an Action nooooo this breaks 'meta: reset_connection' because interpreter_discovery needs an Action
5 years ago
33e7d0804f
fixed tests, and removed the action hack code
5 years ago
7411057ba8
adding in python3 as a possible_pythons option for rhel8
5 years ago
89d87445c8
able to pass tests running interpreter detection in rhel8 container
5 years ago
7d6d76e444
trying to get docker rhel8 working on a Mac host
5 years ago
5e0fb902b2
fix lambda args for fallback
5 years ago
91785f8862
fix another bad import
5 years ago
ee310c1a68
remove_internal_keys returns None, so need different way to check fallback
5 years ago
722cf0ea8d
handle ansible 2.3.3 remove_internal_keys
5 years ago
cf1c3aac1c
fix typo
5 years ago
ba350aa6cf
assigned interpreter vars to the wrong class
5 years ago
d74da84f62
the clean module was added later
5 years ago
496d96c01a
need to convert to str because of unicode_literals
5 years ago
aa0ad5ce8f
accidentally broke discovery
5 years ago
160dbd86b4
keeping track of interpreter discovery vars in case of ansible < 2.8
5 years ago
f10fb7d486
more test cases pass
5 years ago
e3eebb98d0
ImportError instead of newer ModuleNotFoundError
5 years ago
2d3d9cb659
fallback to old-style python interpreter default if running ansible < 2.8.0
5 years ago
ff2fa60184
fixed typo
5 years ago
0b57332d75
all test cases for ansible_python_interpreter on OSX pass now
5 years ago
81694d07ab
fix typo
5 years ago
8630f6190c
fixed propagating discovered interpreter to future tasks; sys.executable still wrong though
5 years ago
af1e22d4e1
able to propagate ansible_python_interpreter fact but still doesn't save between task runs
5 years ago
119e1c24aa
code cleanup
5 years ago
cec692dc17
faking pipelining so discover_interpreter can successfully exit
5 years ago
a6d2b2cd90
doesn't error when declaring 'ansible_python_interpreter: auto' but fact isn't being cached
5 years ago
19ba2417c8
now able to get the action object, but recursion error in low_level_exec_command
5 years ago
806651e112
working on ansible discovery; need to fix passing ActionModuleMixin reference still
5 years ago
d64adb15d7
reverted autolinted code
5 years ago
e8f3154cab
Merge branch 'master' into complexAnsiblePythonInterpreterArg
5 years ago
d6329f3446
Merge devel/290 @ 79b979ec8544ef5d8620c64068d4a42fabf50415
5 years ago
4669c8774f
handles templating ansible_python_interpreter values
5 years ago
cc11864b7d
code cleanup
5 years ago
24b170311a
able to get to 'sudo: source not found' after preventing escape of && so python connects
5 years ago
be4f1bdb50
issue #646 : add extra logging to assertions and start_child()
5 years ago
efd82dd35a
issue #633 : various task_vars fixes
...
- take host_vars from task_vars too
- make missing task_vars a hard error
- update tests to provide stub task_vars
5 years ago
fc09b81949
issue #633 : handle meta: reset_connection when become is active
...
- don't create a new connection during reset if no existing connection
exists
- strip off last hop in connection stack if PlayContext.become is True.
- log a debug message if reset cannot find an existing connection
5 years ago
b6d1df749c
issue #633 : take inventory_hostname from task_vars
...
It used to be set by on_action_run() from task_vars, but this doesn't
work for meta: reset_connection. That meant MITOGEN_CPU_COUNT>1 would
pick the wrong mux to reset the connection on.
5 years ago
f4cf67f0bd
issue #615 : remove meaningless test
...
It has been dead code since at least 2015
5 years ago
e02be89879
issue #625 : ignore SIGINT within MuxProcess
...
Without this, MuxProcess will start dying too early, before Ansible /
TaskQueueManager.cleanup() has a chance to wait on worker processes.
That would allow WorkerProcess to see ECONNREFUSED from the MuxProcess
socket much more easily.
5 years ago
67759371f9
issue #615 : ensure 4GB max_message_size is configured for task workers.
...
This 4GB limit was already set for MuxProcess and inherited by all
descendents including the context running on the target host, but it was
not applied to the WorkerProcess router.
That explains why the error from the ticket is being raised by the
router within the WorkerProcess rather than the router on the original
target.
5 years ago
151b490890
issue #615 : fetch_file() might be called with AnsibleUnicode.
5 years ago
03d2bc6c59
issue #615 : redirect 'fetch' action to 'mitogen_fetch'.
5 years ago
52c8ed7715
issue #615 : extricate slurp brainwrong from mitogen_fetch
5 years ago
069285a588
issue #615 : ansible: import Ansible fetch.py action plug-in
...
From ansible/ansible#9773a1f2896a914d237cb9926e3b5cdc0f004d1a
5 years ago
8dfb3966df
issue #558 , #582 : preserve remote tmpdir if caller did not supply one
...
The undocumented 'tmp' parameter controls whether _execute_module()
would delete anything on 2.3, so mimic that. This means
_execute_remove_stat() calls will not blow away the temp directory,
which broke the unarchive plugin.
5 years ago
3b63da670f
Fix up another handful of LGTM errors.
5 years ago
f78a5f08c6
issue #605 : ansible: share a sem_t instead of a pthread_mutex_t
...
The previous version quite reliably causes worker deadlocks within 10
minutes running:
# 100 times:
- import_playbook: integration/async/runner_one_job.yml
# 100 times:
- import_playbook: integration/module_utils/adjacent_to_playbook.yml
via .ci/soak/mitogen.sh with PLAYBOOK= set to the above playbook.
Attaching to the worker with gdb reveals it in an instruction
immediately following a futex() call, which likely returned EINTR due to
attaching gdb. Examining the pthread_mutex_t state reveals it to be
completely unlocked.
pthread_mutex_t on Linux should have zero trouble living in shmem, so
it's not clear how this deadlock is happening. Meanwhile POSIX
semaphores are explicitly designed for cross-process use and have a
completely different internal implementation, so try those instead. 1
hour of soaking reveals no deadlock.
This is about avoiding managing a lockable temporary file on disk to
contain our counter, and somehow communicating a reference to it into
subprocesses (despite the subprocess module closing inherited fds, etc),
somehow deleting it reliably at exit, and somehow avoiding concurrent
Ansible runs stepping on the same file. For now ctypes is still less
pain.
A final possibility would be to abandon a shared counter and instead
pick a CPU based on the hash of e.g. the new child's process ID. That
would likely balance equally well, and might be worth exploring when
making this code work on BSD.
5 years ago
5af6c9b26f
issue #615 : use FileService for target->controll file transfers
5 years ago
6f12980611
[linear2] merge fallout: re-enable _send_module_forwards().
5 years ago
5298e87548
Split out and make readable more log messages across both packages
5 years ago
0f23a90d50
ansible: log affinity assignments
5 years ago
4f051a38a7
ansible: improve docstring
5 years ago
5811909c8d
[linear2] simplify _listener_for_name()
5 years ago
c68dbdd569
ansible: stop relying on SIGTERM to shut down service pool
...
It's no longer necessary, since connection attempts are no longer truly
blocking. When CTRL+C is hit in the top-level process, broker will begin
shutdown, which will cancel all pending connection attempts, causing
pool threads to wake. The pool can't block during shutdown anymore.
5 years ago
f4ca926b21
ansible: cleanup various docstrings
5 years ago
edde251d58
issue #549 : ansible: reduce risk by capping RLIM_INFINITY
5 years ago
Steven Robertson