Merge lp://staging/~simpoir/landscape-client/fix_jenkins_tests_taking_too_long into lp://staging/~landscape/landscape-client/trunk
Proposed by
Simon Poirier
Status: | Merged | ||||
---|---|---|---|---|---|
Approved by: | Simon Poirier | ||||
Approved revision: | 919 | ||||
Merged at revision: | 918 | ||||
Proposed branch: | lp://staging/~simpoir/landscape-client/fix_jenkins_tests_taking_too_long | ||||
Merge into: | lp://staging/~landscape/landscape-client/trunk | ||||
Prerequisite: | lp://staging/~simpoir/landscape-client/fix_tempfile_cleaning | ||||
Diff against target: |
15 lines (+3/-1) 1 file modified
landscape/tests/test_watchdog.py (+3/-1) |
||||
To merge this branch: | bzr merge lp://staging/~simpoir/landscape-client/fix_jenkins_tests_taking_too_long | ||||
Related bugs: |
|
Reviewer | Review Type | Date Requested | Status |
---|---|---|---|
Chad Smith | Approve | ||
Geoff Teale (community) | Approve | ||
🤖 Landscape Builder | test results | Approve | |
Review via email: mp+303314@code.staging.launchpad.net |
Commit message
bump max-retries for watchdog tests to avoid false negatives on slow test
setup on some jenkins slaves.
Description of the change
bump max-retries for watchdog tests to avoid false negatives on slow test
setup on some jenkins slaves.
Testing instructions:
This one is tricky to reproduce, as running the tests on the failing nodes
multiple times will make it succeed on the subsequent runs, until kernel
caches expires between runs (like scheduled jenkins runs).
I had some success reproducing by sync && echo 3 > /proc/sys/
before a run and adding an io limit to an lxd container, so it runs as slow
as jenkins machines.
To post a comment you must log in.
Command: TRIAL_ARGS=-j4 make check /ci.lscape. net/job/ latch-test/ 7875/
Result: Success
Revno: 919
Branch: lp:~simpoir/landscape-client/fix_jenkins_tests_taking_too_long
Jenkins: https:/