Repository navigation
[BUG] sof-logger resisted pkill -INT, using -KILL #385
Description
Activity
Quick workaround submitted at #381
- added a commit that references this issue
on Sep 18, 2020 I cannot reproduce locally but I got an idea:
2020-09-17 20:56:25 UTC [REMOTE_ERROR] sof-logger resisted pkill -INT, using -KILLI don't do anything
REMOTE_maybe that's why it's happening in CI and not locally.From https://pubs.opengroup.org/onlinepubs/7908799/xsh/signal.h.html
SIGINT | i | Terminal interrupt signal.
I've never done the REMOTE_ thing, no idea how to do it. @xiulipan or @aiChaoSONG could you compare REMOTE_ vs local for me?
I switched from -INT to -TERM and it works, please review: #382
I filed this in sof because I was afraid it was an issue with the logger but now I think it's just an sof-test issue.
Keep this issue open before we root cause if this is a bug of sof-logger or test case.
- If this is an sof-logger issue, will close this issue and report a bug in SOF repo.
- If this is a test case issue, we will use PR refine alsabat test script #401 and close this issue.
- addedtype:bugSomething doesn't work as expectedSomething doesn't work as expectedP2Critical bugs or normal featuresCritical bugs or normal features
on Sep 28, 2020 I believe this has been fixed by #382
hijack.sh: switch first attempt to stop the logger from -INT to -TERM, closing.- addedarea:logsLog and results collection, storage, etc.Log and results collection, storage, etc.
on Jul 3, 2021
#373 was merged because https://sof-ci.01.org/softestpr/PR373/build182/devicetest was all green. However almost all tests were failing because the sof-logger resisted kill -INT. To make it worse I misused
exit_status=2which has the special meaning Not Applicable which means no error and everything is green.