I've started to see consistent failures with the client plugin e2e's: A snippet from the debug jenkins failure dump: [Pipeline] echo in catch block [Pipeline] echo Caught: hudson.AbortException: create returned an error; {reference={}, err=Error from server (AlreadyExists): error when creating "https://raw.githubusercontent.com/openshift/nodejs-ex/master/openshift/templates/nodejs.json": templates.template.openshift.io "nodejs-example" already exists , verb=create, cmd=oc --server=https://172.30.0.1:443 --certificate-authority=/var/run/secrets/kubernetes.io/serviceaccount/ca.crt --namespace=e2e-test-jenkins-pipeline-kbdph --token=XXXXX create -f https://raw.githubusercontent.com/openshift/nodejs-ex/master/openshift/templates/nodejs.json -o=name , out=, status=1} [Pipeline] End of Pipeline ERROR: create returned an error; {reference={}, err=Error from server (AlreadyExists): error when creating "https://raw.githubusercontent.com/openshift/nodejs-ex/master/openshift/templates/nodejs.json": templates.template.openshift.io "nodejs-example" already exists , verb=create, cmd=oc --server=https://172.30.0.1:443 --certificate-authority=/var/run/secrets/kubernetes.io/serviceaccount/ca.crt --namespace=e2e-test-jenkins-pipeline-kbdph --token=XXXXX create -f https://raw.githubusercontent.com/openshift/nodejs-ex/master/openshift/templates/nodejs.json -o=name , out=, status=1} Finished: FAILURE Instances of this: - https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/pr-logs/pull/openshift_openshift-controller-manager/59/pull-ci-openshift-openshift-controller-manager-release-4.3-e2e-aws-proxy/16 - https://prow.svc.ci.openshift.org/view/gcs/origin-ci-test/pr-logs/pull/openshift_openshift-controller-manager/59/pull-ci-openshift-openshift-controller-manager-release-4.3-e2e-aws-builds/14
@Steve - as this is a test case fix I'm good with it getting in sooner rather than later. So if targeting to 4.3.0 enables that, great. But of course if push comes to shove unless the test case problem is blocking an actual 4.3.0 GA fix, this should not block 4.3.0. Given that, if targeting to 4.3.z makes more sense, please let us know.
OK Steve and I talked in the hallway. Since the jenkins job is "optional" wrt e2e jobs monitored for 4.3.0 GA, we can mark this 4.3.z This only occurs in 4.3 currently because the 4.4 samples are updated to deal with nodejs EOL images in the nodejs imagestream. The fix here is either a) undo https://github.com/openshift/jenkins-client-plugin/pull/317 b) or update that test to overrider the version parameter in the template and use the older nodejs version still available on 4.3 c) IF we decide to backport the nodejs EOL changes to 4.3 via https://bugzilla.redhat.com/show_bug.cgi?id=1788065 this should work on 4.3.z as well ... but that is not a given at this point This bug is open to track a) or b) ... note, only a jenkins-client-plugin master branch PR is needed for a) and b) and updating the test shared across releases and we do not have to go through the standard 4.x z stream cherry pick approval process for this test change.
Sorry I may have gotten my client plugin PRS and nodejs in flight items confused I is possible that https://github.com/openshift/jenkins-client-plugin/pull/318 may have fixed this issue testing in https://github.com/openshift/openshift-controller-manager/pull/59 now ... if the e2e passes there, will close out if not, I'll analyze, but I suspect the versioning situation noted in #Comment 3 may then take hold reporting back in a bit