action #38702
closed
[functional][u] apply workaround for bsc#1096401 "snapper terminates with core dump"
Added by okurz almost 6 years ago.
Updated almost 6 years ago.
Category:
Bugs in existing tests
Target version:
SUSE QA - Milestone 18
Description
Motivation¶
Under the assumption that bsc#1096401 is actually a symptom of other random failures in random test modules we should detect the bug presence, record a soft failure and apply a workaround, e.g. disable snapper
Acceptance criteria¶
- AC1: soft_failure to bug is raised when detected
- Copied from action #36304: [opensuse][functional]sporadic][u] chromium test is unstable added
- Copied from deleted (action #36304: [opensuse][functional]sporadic][u] chromium test is unstable)
- Blocks action #36304: [opensuse][functional]sporadic][u] chromium test is unstable added
- Subject changed from [functional][u] apply workaround for bsc#1079877 to [functional][u] apply workaround for bsc#1096401 "snapper terminates with core dump"
- Description updated (diff)
- Status changed from New to Workable
- Priority changed from Normal to Urgent
- Related to action #38387: [functional][u][sporadic][medium] test fails in gnucash - window is not closed added
- Blocks coordination #35302: [qe-core][opensuse][functional][epic][sporadic] Various unstable tests on o3 added
- Blocks action #38834: [sle][functional][y] Verify bsc#1096401 using openQA added
- Status changed from Workable to Resolved
- Assignee set to okurz
The patch has reached openSUSE Tumbleweed but at least some tests still fail:
[27 Jul 2018 09:43:40] <DimStar> okurz: so, https://openqa.opensuse.org/tests/714806 still fails VLC (still that lxde, which is fine on DVD; will be a fun one) - and the journal no longer shows coredump generation for snapperd
[27 Jul 2018 10:12:15] <okurz> DimStar: thx for letting me know. This is what I expected that snapper core dumping during its shutdown is not really the issue but we ruled it out, that's a start :)
[27 Jul 2018 10:12:58] <DimStar> okurz: well, we did expect it to spoil disk i/o in some scenaios, so having that noise in the logs gone is definitively a good thing
[27 Jul 2018 10:13:42] <okurz> DimStar: sure. I also did not rule out any further I/O causing tests to fail but the LVC looks more reproducible with exactly the same screen
[27 Jul 2018 10:14:30] <DimStar> okurz: VLC now only failed in one scenario at least (the lxde/NET)
With the patch in the product we do not need a workaround anymore and can continue in #38834 to verify it can not happen again unless we just accept it as good enough :)
Also available in: Atom
PDF