| ssastry | |
| May 12 2025, 3:00 PM |
| F59932991: image.png | |
| May 13 2025, 10:14 AM |
It happens not so infrequently that testreduce1002 stalls for short or long periods of time when running rt-tests. I haven't investigated but I think this is because of processes thrashing which might be mitigated by adding some more RAM to this VM.
What is involved in bumping up RAM on this VM by another GB?
It would seem you're right about memory pressure being an issue:
It involves rebooting the VM, there's enough RAM on the host (and generally in the cluster). Tell us when we can do that without disruption
for you and we will.
Anytime today or tomorrow works. We'll hold off running rt-testing till the reboot happens.
Mentioned in SAL (#wikimedia-operations) [2025-05-13T15:56:21Z] <claime> gnt-instance modify -B memory=10g testreduce1002.eqiad.wmnet - T393904
VM testreduce1002.eqiad.wmnet rebooted by cgoubert@cumin1002 with reason: Pick up new 10GB ram
cgoubert@testreduce1002:~$ free -m
total used free shared buff/cache available
Mem: 9944 2211 7646 0 323 7733
Swap: 975 0 975Done :)