It looks like osci-jenkins-1 is having trouble creating pods: https://osci-jenkins-1.ci.fedoraproject.org/job/eln-build-pipeline/47890/console
Many (all?) builds seem to be affected.
Metadata Update from @msrb: - Issue assigned to astepano
@astepano could you please take a look what is going on? thanks :wink:
I think this is because Java on master process eats all available resources. Jenkins fails check jenkins-runners. We didn't find out why Java process is so high. We have a new VPC in new AWS region. Currently I cannot scale K8S resources. Need to wait until we move to new VPC + make some optimization to existing configuration.
Note the problem is that kubernetes plugin is unable to create any pods. Currently there is 0 pods attached to Jenkins master. This started on Friday (evening I think). This was not a problem before.
And just to be clear: since no pods can be created, ELN + generic + functional pipelines will all either block (and builds pile up, waiting indefinitely), or timeout after couple of hours. So the Jenkins instance is basically unusable at the moment.
I believe we found cause for this. I sent PR/merged ELN pipelines. The reason for this 40K + 60K history length + EBS as storage backend for k8s nodes.
Metadata Update from @astepano: - Issue status updated to: Closed (was: Open)