* minor service_mgr facts fixes
handle case in which ps command fails or returns empty
updated tests since it now does keep trying to detect after ps fails
cli options will now display either use provided info, or automating from the name
Co-authored-by: Sloane Hertel <19572925+s-hertel@users.noreply.github.com>
- remove need for module (at least for posix side)
- adds retry with backoff on fetching file, since
race is bigger since we don't spend time on module
- now gives more info on fail
- also made actionfail/skip handle results if given
* rebased with upstream
* removed extra usetty as it wasnt needed, style changes, added var option setable by inventory for pkcs11
* update pkcs11_provider version_added
* Update lib/ansible/plugins/connection/ssh.py
Co-authored-by: Jordan Borean <jborean93@gmail.com>
* Update lib/ansible/plugins/connection/ssh.py
Correct logic for a password being required for pkcs11_provider
Co-authored-by: Jordan Borean <jborean93@gmail.com>
* style nit fixes for checking pkcs11_provider is set
* fixed duplication when using password_prompt with pkcs11_provider
* added changelog fragment
* added changelog fragment
Co-authored-by: David Whiteside <david.whiteside@nrel.gov>
Co-authored-by: Matt Davis <nitzmahone@users.noreply.github.com>
Co-authored-by: Jordan Borean <jborean93@gmail.com>
base._update_security_filters is a private attribute of DNF used
as performance optimization. Modification or even call from outside
of DNF is against all recommendation including PEP8.
* Improve compatibility with all DNF versions
* Add changelog fragment for dnf security change
* Add resiliency to linux hw fact gathering
Now traps unexpected exceptions on mounts and continues
gathiering other info.
Also gives more info on why mount info gathering failed.
more info if debugging
Changed confusing wording. People new to ansible, like myself, have a hard time grocking what is a playbook and what is a task. I changed the name of the playbook to make it clearer.
* Due to the takeover of freenode we're moving to a different irc network.
* Our channels updated to point at the same channel name on libera.chat
* Some links went to webchat.freenode.net. At this time, libera.chat
doesn't point you to an official webchat client so I changed these to
https://libera.chat. (kiwi irc does work with libera.chat so that
could be another option).
* In general, I used the name irc.libera.net for link names and
https://libera.chat for link targets. This is because the irc service
is hosted on irc.libera.chat but the project web server is hosted on
libera.chat. (This appears to also be true for freenode but we were
using http://irc.freenode.net which doesn't seem to work. Oops).
* Removed http://irc.freenode.net from the linkcheck exceptions.
linkcheck was actually correct to flag that as invalid (should have
been http://frenode.net instead).
* Looks like hte important people in #yaml are now in libera.chat
* Link to where contributors should get help
Add a link target and then link to where contributors should get support
for developing groups of modules.
* Update docs/docsite/rst/dev_guide/developing_modules_in_groups.rst
Co-authored-by: Felix Fontein <felix@fontein.de>
Co-authored-by: John R Barker <john@johnrbarker.com>
Co-authored-by: Felix Fontein <felix@fontein.de>
* Fix module-specific defaults in the gather_facts, package, and service action plugins.
* Handle ansible.legacy actions better in get_action_args_with_defaults
* Add tests for each action plugin
* Changelog
Fixes#72918
Change:
- Make strategies behave consistently and return the empty string
instead of "UNKNOWN" (or "temporarystub") for the "before" value if
the permanent hostname file does not exist or could not be read.
- Switch to `with open()` instead of annoying exception handling code
(which was wrong and leaked file handles in several places). This
drops Python 2.4 support for this module.
- Updated porting guide since users could be relying on these former,
inconsistent values.
Signed-off-by: Rick Elrod <rick@elrod.me>
Change:
- Hostname strategies' get_*() methods should never write to the
filesystem. They are used in check_mode by default to determine if
there is any work to be done.
Test Plan:
- New unit tests to ensure that (at least when in check_mode) the get
methods don't ever call write.
Tickets:
- Fixes#66432
Signed-off-by: Rick Elrod <rick@elrod.me>
- AWS has moved to Zuul CI
- Kubernetes collection has been renamed
- OpenShift/OKD collection was missing
- ansible.windows is on Zuul and Azure Pipelines
- ansible.posix is on Zuul
Signed-off-by: Jill Rouleau <jill.rouleau@bespokess.com>
Signed-off-by: Abhijeet Kasurde <akasurde@redhat.com>
