On Thu, 6 Apr 2023 at 07:57, Thomas Huth <thuth@redhat.com> wrote:
On 05/04/2023 17.15, Peter Maydell wrote:
The s390 private runner CI job ubuntu-20.04-s390x-all seems to have
started timing out a lot recently. Here's an example where it passed,
but with only 53 seconds left on the clock before it would have been
killed:
https://gitlab.com/qemu-project/qemu/-/jobs/4066136770
It looks like 'make check' was about 30 minutes of the 75 minutes
total, and compilation was 45 minutes.
Any suggestions for how we can trim this down? (Presumably we
could also raise the time limit given that this is a private
runner job...)
I don't have access to that system, so I can only guess: Did you check
whether there are any other processes still running (leftovers from earlier
test runs)?
I did check for that, as it's been a problem in the past, but
no, in this case no other jobs are running in the VM.
If not, it's maybe because it is a VM that is running with other
VMs in parallel that hog the CPU? In that case, you could contact the owner
of the machine and ask whether there is anything that could be done about
it. Or simply increase the timeout a little bit more... (our highest timeout
in another job is 90 minutes, so I think that would still be OK?).
Christian, does our S390X machine get a guaranteed amount of CPU,
or does it depend on what else is running on the hardware?