Release Manager: issue when Release Manager pipeline try to retrieve a existing pod

[jssnull]

When deploying an RShiny app in QA, the app gets deployed successfully in OpenShift by the creation of a ReplicationController of 1 replica and the Pod gets deployed successfully and it's up and running. However, the Jenkins pipeline is marked as failed. The reason for this is that there's a step where it tries to find the Pod that got created and it cannot find it. We have the suspicion that it could be related to a timing issue - when it tries to find the Pod, the Pod has not been created as such in the cluster.

This would be the error reported by jenkins:

We can see that the pod was successfully created, but wasn’t recognized in the last log:

We can see the final status of the pipeline graphically when it fails:

This issue happens when we try to deploy to qa and prod, If we try to launch the pipeline again it will work, but wouldn’t create a new pod. This issue could be produced by a timinig problem.

Maybe a sleep function could be added here to solve the issue: https://github.com/opendevstack/ods-mro-jenkins-shared-library/blob/2.x/src/org/ods/service/OpenShiftService.groovy#L280-L310

In ODS 3.x there is a similar sleep function implemented, it could be helpful. https://github.com/opendevstack/ods-jenkins-shared-library/blob/3.x/src/org/ods/services/OpenShiftService.groovy#L433-L459

[segator]

I think we should update the code to wait for pod readiness before try to extract the information we can use labels for that for exemple o rdeployment reference. even we can use podname but if our deployment object have multiple replicas best to wait for all pods been runing. Something like

oc get pods -l app=hello -o 'jsonpath={..status.conditions[?(@.type=="Ready")].status}') != "True"

[clemensutschig]

@jssnull - can you try the same with ODS3 .. and see if that solves it?