* Fix types when evaluating interpreter (#36541)
* Fix types when evaluating interpreter. Fixes#36536
* Rename variables that contain bytes to b_*
* Get rid of to_text() and to_bytes() calls that do nothing (because the
data is already the proper type)
(cherry picked from commit a4df4d33ac)
* Changelog entry for the ansible_*_interpreter fix
* Add example on how to use a here document with shell module
* Remove here doc example and add note to use script module instead
(cherry picked from commit 5e7ee9df0a)
* If inventory file isn't able to be parsed by aws_ec2, return an empty dict instead of None
* Raise an AnsibleParserError instead
* remove extra lines
* aws_ec2 inventory plugin - fix path matching logic
* Re-use logic from StrategyBase._load_included_file in StrategyModule.run for free and linear (#36470)
This improves include_role performance and recursion limits
(cherry picked from commit 10fefc7156)
* Add changelog for 36470
* Better handling of absent AWS SES identity notification information.
Fixes#36065
aws_ses_identity module now handles the cases where information about
the notification setup for the identity isn't returned by the AWS api.
This seems to happen in an edge case, believed to be eventual
consistency on registering new identities. So this case is treated
as if has been no notification setup for the identity yet.
Also fix 2 flake8 warnings in the module, a missing newline and unused
import.
* Increase the Boto Retries on SES APIs to deal with throttling.
This should address the unstable integration test failing due to
parallel runs in shippable hitting AWS throttling.
* Add retries loading SES details for inclusion in successful response.
There seems to be an eventual consistency behaviour with identity
registration. It's possible to still get no identity back after
registration.
This can cause failures in the shippable builds. This should fix that by
creating a retry of retrieving the identity information after
registration.
A similar retry loop has been added to notification attributes to ensure
this doesn't suffer from the same failure.
* Add missing sleep in get_notification_attributes to avoid busy loop.
* Add force option to systemd module
* Use multi-line YAML syntax in examples
* Add version_added for new option
(cherry picked from commit 2bffcfa63b)
* fetch: set fail_on_missing: True as default as per docs
* Updated docs for fetch to say behaviour was changed in 2.5 and updated tests
(cherry picked from commit df8a5d7a4f)
* Bug in del(list) logic. Say you have a list of 4 elements a[0-3]
and you have to remove index 1 and 3, if you remove index 1 first
then list is cut short a[0-2] and does not have index 3
Fix: Remove indexes in reverse sorted order e.g. 3 and 1 in above
example so that order of indexes remain preserved even after deleting
fix is to remove indexes in reverse sorted order
* Add test cases for failed case
(cherry picked from commit 0bbea9a579)
* now get_url and other modules default to module temp dir
also fixed 'bare' exception
* allow modules to work with older versions
* updated docs per feedback
(cherry picked from commit c119d54e4a)
Unified tmp accidentally removed the containing tmpdir from the list of
files to fix the permissions on when we're becoming a different
unprivileged user. This resulted in a visible bug for script but not
for patch. This is because patch also uploads the module to the same
temporary directory and the uploaded module also ends up calling
fixup_perms2() which includes the temporary directory. So by the time
patch needs to access the temporary patch file, the directory is
appropriately set.
script's breakage was visible because script does not upload a module
(it's akin to raw in this way). Therefore, we only call fixup_perms2()
once in script and so leaving out the tmpdir in script means that the
containing directory never has its permissions set appropriately.
Fixing both because it does not cause an extra round trip for patch so
any speedup would be minimal and it's better to fix the perms as close
as possible to where we know we need it. Otherwise, changes to
seemingly unrelated code later could end up breaking it.
Fixes#36398
(cherry picked from commit edaeb69a35)
* Update network debug troubleshooting guide
Fix#35914
Command timeout and connection timeout error messages
are dsiplayed in log file instead on console.
Update the same in troubleshooting guide.
* Update example error
* Fix CI issues
* Fix more CI failures
* More fixes
* Fix review comments
* Fix more review comments
* Copy edit
(cherry picked from commit b57cc7cf31)
* Fix nxos_system tests (#36201)
* Fix nxos_system tests
* Add debug connection plugin info
* Move sanity test under common
(cherry picked from commit 5b5d24631a)
* Fixes for the N3500 platform that uses the A8 image (#36261)
* fix nxos_l3_interface tests as n35 doesn't support ipv6
* add terminal dont-ask to nxos_feature and nxos_lldp
* put interfaces in L2 mode for N35
* fix nxos_feature unit-tests
(cherry picked from commit e24c547a3a)
send_command already performs the to_bytes safely on prompts (checking
for None). Without this check the literal 'None' became a subprompt trigger!
Fixes#35662
(cherry picked from commit f2818caa61)
ansible-console was calling setup_vault_secrets
twice. Once directly and once via _play_prereqs()
The direct invocation was not setting auto_prompt=False.
However, the direct invocation isn't need at all so
this removes it so only _play_reqs() is used.
That fixes the unrequested vault password
prompting.
Fixes#33027
(cherry picked from commit 810fa7046b)
Ryan Brown