kube-apiserver received an error while watching events: The resourceVersion for the provided watch is too old. Failing tests: [Disruptive] Cluster upgrade should maintain a functioning cluster [Feature:ClusterUpgrade] [Suite:openshift] [Serial] https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-aws-upgrade-4.0/1017 https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-aws-upgrade-4.0/1016 https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/logs/release-openshift-origin-installer-e2e-aws-upgrade-4.0/1015 https://openshift-gce-devel.appspot.com/build/origin-ci-test/logs/release-openshift-origin-installer-e2e-aws-upgrade-4.0/1017/
Apr 09 11:57:01.789 W ns/openshift-marketplace pod/community-operators-8dfb4dc87-pfd6z Readiness probe errored: rpc error: code = Unknown desc = command error: command timed out, stdout: , stderr: , exit code -1 (5 times)
https://storage.googleapis.com/origin-ci-test/logs/release-openshift-origin-installer-e2e-aws-upgrade-4.0/1017/artifacts/e2e-aws-upgrade/pods/openshift-marketplace_community-operators-6c959f865-4sbzw_community-operators.log.gz The pod is successfully starting, but takes about 45s to do so. Can the test be adjusted to allow for this?
Ok, sounds like the slow community-operators is a red-herring. There were no failing cluster-operators here: $ curl -s https://storage.googleapis.com/origin-ci-test/logs/release-openshift-origin-installer-e2e-aws-upgrade-4.0/1017/artifacts/e2e-aws-upgrade/clusteroperators.json | jq '.items[] | .failing = ([.status.conditions[] | select(.type == "Failing")][0]) | .pending = ([.status.conditions[] | select(.type == "Pending")][0]) | select(.failing.status == "True" or .pending.status == "True")' But the version conversion was not complete, which is probably why the test failed: $ curl -s https://storage.googleapis.com/origin-ci-test/logs/release-openshift-origin-installer-e2e-aws-upgrade-4.0/1017/artifacts/e2e-aws-upgrade/clusteroperators.json | jq -r '.items[] | .version = (.status.versions[] | select(.name == "operator").version) | .version + " " + .metadata.name' | sort 4.0.0-0.ci-2019-04-09-033744 authentication 4.0.0-0.ci-2019-04-09-033744 cluster-autoscaler 4.0.0-0.ci-2019-04-09-033744 console 4.0.0-0.ci-2019-04-09-033744 image-registry 4.0.0-0.ci-2019-04-09-033744 ingress 4.0.0-0.ci-2019-04-09-033744 marketplace 4.0.0-0.ci-2019-04-09-033744 monitoring 4.0.0-0.ci-2019-04-09-033744 node-tuning 4.0.0-0.ci-2019-04-09-033744 openshift-controller-manager 4.0.0-0.ci-2019-04-09-033744 openshift-samples 4.0.0-0.ci-2019-04-09-033744 service-catalog-apiserver 4.0.0-0.ci-2019-04-09-033744 service-catalog-controller-manager 4.0.0-0.ci-2019-04-09-033744 storage 4.0.0-0.ci-2019-04-09-111643 cloud-credential 4.0.0-0.ci-2019-04-09-111643 dns 4.0.0-0.ci-2019-04-09-111643 kube-apiserver 4.0.0-0.ci-2019-04-09-111643 kube-controller-manager 4.0.0-0.ci-2019-04-09-111643 kube-scheduler 4.0.0-0.ci-2019-04-09-111643 machine-api 4.0.0-0.ci-2019-04-09-111643 machine-config 4.0.0-0.ci-2019-04-09-111643 network 4.0.0-0.ci-2019-04-09-111643 openshift-apiserver 4.0.0-0.ci-2019-04-09-111643 operator-lifecycle-manager 4.0.0-0.ci-2019-04-09-111643 operator-lifecycle-manager-catalog 4.0.0-0.ci-2019-04-09-111643 service-ca I think that is just bug 1694226, so I'm closing this as a dup. *** This bug has been marked as a duplicate of bug 1694226 ***