Comment 35 for bug 1640547

Revision history for this message
Rod Smith (rodsmith) wrote :

Thanks for the work on this so far, Colin. One point:

> I'm still not happy about the /usr/lib/plainbox-provider-checkbox/bin/disk_stress_ng using timeout with a -9 (SIGKILL) to terminate stress-ng stressors. Stress-ng stressors can be *cleanly* terminated with a SIGALRM signal, this triggers all the processes to terminate once they have freed resources.

We tried this, and unfortunately, SIGALRM didn't work; it doesn't terminate the test in the case of some failures (the one described in this bug report, for instance). For our purposes, if one test fails, the test as a whole has failed, so we aren't too concerned with any subsequent failures that might be a result of using SIGKILL to terminate the first failure. We ARE concerned, though, with the test suite hanging, which could take hours to discover. If the disk corruption affects subsequent tests, then that is of course a concern, but so far we haven't encountered a problem with that.