2025.1 sapcc cherry picks and required changes

[stanislav-zaprudskiy]

The previous changeset stable/2024.1...sapcc:keystone:stable/2024.1-m3

diff command (collapsible) 👇

$ diff \
  <(git log --no-decorate --oneline origin/stable/2024.1...origin/stable/2024.1-m3 | cut -d ' ' -f 2-) \
  <(git log --no-decorate --oneline origin/stable/2025.1...2025.1-cherry-picks | cut -d ' ' -f 2-)
1,3d0
< Merge pull request #35 from sapcc/cherry_pick_ossa-2025-002
< Add service user authentication to ec2 and s3 endpoints
< Add test with noauth for s3tokens and ec2tokens
5d1
< Merge pull request #33 from sapcc/bmemcached
7d2
< Merge pull request #31 from sapcc/chery-pick-pci-dss-invalid-password-reporting
9d3
< Merge pull request #29 from bbobrov/stable/2024.1-m3
15d8
< Correct format for token expiration time
21d13
< Catch CredentialNotFound and raise a proper exception

lost the following commits:

• 98b29c6, a5ce512, 8d7b141 as these are now part of the upstream (0d46250, 90695d8)
• 0273db1, as merge commit of 3aaa4fd (which is included)
• 04df531, as merge commit of c5641f4 (which is included)
• 72b4f60, as merge commit of 6e82d1f, cbcb56a (which are included)
• 92d2c28, as it is now part of the upstream (d01cde5)
• 1b013fd (originally added in Catch a couple uncaught exceptions #22), as it's been superseded by upstream's ae4e541
• 3c93bbe, as not relevant
• dc9491a and cf41a74, as mutually exclusive and not relevant

Additions:

• Commit pep8 suggestions
• Switch dependencies to 2025.1
• concourse_unit_test_task: use py312 - as loci image for epoxy is Ubuntu Noble based, which comes with python3.12
• Rework concourse_unit_test_task

[stanislav-zaprudskiy]

I was able to execute py312 tests locally

Ran: 5723 tests in 154.6647 sec.
 - Passed: 5006
 - Skipped: 715
 - Expected Fail: 2
 - Unexpected Success: 0
 - Failed: 0

[JoJoPuppe]

I checked how you arrived at the proposed cherry-picked commits and your reasoning behind why certain commits are no longer needed. It makes sense to me, and running the test locally with py312 resulted in nearly the same outcome as yours. Somehow, one more test was run, but it was also successful.

======
Totals
======
Ran: 5724 tests in 94.0914 sec.
 - Passed: 5007
 - Skipped: 715
 - Expected Fail: 2
 - Unexpected Success: 0
 - Failed: 0
Sum of execute time for each test: 1029.0674 sec.

[tz3]

Thanks for the PR.

I ran through the commits LGTM could run the test locally (with docker)

Totals
======
Ran: 3269 tests in 109.8747 sec.
 - Passed: 2868
 - Skipped: 400
 - Expected Fail: 1
 - Unexpected Success: 0
 - Failed: 0
Sum of execute time for each test: 861.7773 sec.

==============
Worker Balance
==============
 - Worker 0 (407 tests) => 0:01:49.812731
 - Worker 1 (407 tests) => 0:01:49.209030
 - Worker 2 (409 tests) => 0:01:49.016440
 - Worker 3 (409 tests) => 0:01:48.744222
 - WARNING: missing Worker 4!
 - WARNING: missing Worker 5!
 - WARNING: missing Worker 6!
 - WARNING: missing Worker 7!
 - Worker 8 (409 tests) => 0:01:47.475433
 - Worker 9 (409 tests) => 0:01:45.047961
 - WARNING: missing Worker 10!
 - WARNING: missing Worker 11!
 - Worker 12 (409 tests) => 0:01:45.182973
 - Worker 13 (410 tests) => 0:01:49.447896
  py312: OK (208.53=setup[80.58]+cmd[127.95] seconds)
  congratulations :) (208.82 seconds)```