Cancelling the running of Jenkins jobs shouldn't cause the system to fail trying to provision a pod/vm in a way that we see a continual failure/retry repeating until some time period before we get a pod that we can use.
Scott
On Thu, Nov 10, 2022, 12:39 AM Ed Willink <ed@xxxxxxxxxxxxx> wrote:
Hi
This might be 'my fault'. I managed to do the wrong 'Push' as a
result of which 150 Gerrit jobs were queued for OCL. It took me
some time to notice and even longer to kill them all, clicking on
the red cross one at a time. There seems to be no multi-select.
Regards
Ed Willink
On 10/11/2022 02:34, Jonah Graham
wrote:
Hi Denis,
From my perspective things are looking a little unusual,
not sure exactly why. Is the cluster overloaded?
For example this build took 1h30 to provision
a pod, and like Scott's messages, it looks like some
failure/retry is happening because it looks like lots of
different pods are being created (notice in the log I linked
to that each "Created Pod" line is different). It's not just
CDT, as LSP4J had a ~20 minute delay with
6 new "Created Pod" lines too.