backport some skips from main - #23219
Conversation
Many are fixable in priciple, but let's do this in a series of follow-ups. We want to establish a baseline of RHEL 8 tests first and not regress further. Mark these with "FIXME". Some tests, e.g. Stratis, will never work on RHEL 8 any more, as we removed the code support for that. Skip them with a precise reason.
RHEL 8.10's cryptsetup uses too much memory (https://issues.redhat.com/browse/RHEL-8258). So the LUKS tests are too flaky in RHEL 8. Skip them, like in check-storage-resize.
martinpitt
left a comment
martinpitt
left a comment
There was a problem hiding this comment.
When backporting, please always add a "cherry-picked from main commit 123abc" to the commit message, to make these easier to track. Thanks!
| self.login_and_go("/metrics") | ||
|
|
||
|
|
||
| @testlib.skipImage("FIXME: python3-pcp not currently on image", "rhel-8-10") |
There was a problem hiding this comment.
I don't understand these. The last rhel-8 landed PR #23138 was happy. Either we added python3-pcp to the rhel-8-10 image already, or something else is wrong and broke the rhel-8 branch on main?
|
|
||
|
|
||
| @testlib.nondestructive | ||
| @testlib.skipImage("cryptsetup uses too much memory, OOM on our test VMs", "rhel-8-*") |
There was a problem hiding this comment.
I thought we fixed these on main by running nond machines with more memory. t/c/run-tests has
# on RHEL 8, provide enough RAM for cryptsetup's PBKDF (https://issues.redhat.com/browse/RHEL-8258)
# On any newer OS, cryptsetup gets along with less RAM
if image.startswith("rhel-8") and opts.nondestructive_memory_mb == DEFAULT_MACHINE_MEMORY_MB:
opts.nondestructive_memory_mb = 1400
Perhaps we should backport that rather?
There was a problem hiding this comment.
Judgement call, and I trust your judgement :)
|
What the heck happened to the test triggers here? This is only supposed to run rhel-8.. |
It's because I accidentally opened the branch against main (which was preselected for some reason, instead of rhel-8). When I rebased the branch, GitHub cleaned up the PR UI but since we report statuses against HEAD commits, they stayed around... |
|
My backport to RHEL 8 went completely green #23026 so I am wondering about the reason this PR exists. |
3 participants
Thee got added to main instead of rhel-8, so they have no effect on the normal rhel-8 runs (vs. the rhel-8 wscontainer scenario which does run from main). Backport them to the rhel-8 branch.