Closed shangsong2 closed 1 month ago
Hello. Please do read and follow the PR template, it makes it easier for me to understand what you are trying to do and for you to explain it.
As for your change, if what you are trying to do is increase this timeout, this is not the right way to go about it. That 10% is ment to be a small grace period, not the way we regulate the timeout. Why aren't you using the parameter to customize the timeout?
Closing it as there was no activity, feel free to reopen it
For a large configuration server with 480 cores and 2TiB memory, it need over 2 times time to finish the CPU stress test. For example, below value of timeout is 900 seconds, but stress-ng actually takes over 30 minutes(1807.05 secs) to complete(10% runtime is not enough). $ sudo /usr/lib/checkbox-provider-base/bin/stress_ng_test.py cpu -b 900 stress-ng: info: [139734] load average: 535.37 989.84 2322.35 stress-ng: info: [139734] skipped: 0 stress-ng: info: [139734] passed: 7199: af-alg (479) bsearch (480) context (480) cpu (480) crypt (480) hsearch (480) longjmp (480) lsearch (480) matrix (480) qsort (480) str (480) stream (480) tsearch (480) vecmath (480) wcs (480) stress-ng: info: [139734] failed: 0 stress-ng: info: [139734] metrics untrustworthy: 0 stress-ng: info: [139734] successful run completed in 30 mins, 7.05 secs
retval is 0
So it is better to increase the timeout from 10% to 300%.
Description
Resolved issues
Documentation
Tests