Upgrade process does not complete
After you upgrade IBM Cloud Pak foundational services, the replica count of the IBM Cloud Pak foundational services
operator or the operand-deployment-lifecycle-manager
operator is zero.
Symptom
The upgrade process does not complete.
-
Verify the replica count of the
IBM Cloud Pak foundational services
operator:oc get deployment ibm-common-service-operator -n <namespace-where-foundational services-are-installed> -o=jsonpath='{.status.replicas}'
You might see the following output:
0%
-
Verify the replica count of the
operand-deployment-lifecycle-manager
operator:oc get deployment operand-deployment-lifecycle-manager -n <namespace-where-foundational services-are-installed> -o=jsonpath='{.status.replicas}'
You might see the following output:
0%
Cause
You might see this issue if you rerun the upgrade script before the upgrade completes. The script, when you run it the second time, scales down the replica count of the IBM Cloud Pak foundational services
operator or the operand-deployment-lifecycle-manager
operator to 0. The script also modifies the common-service
OperandRegistry.
Resolving the problem
To resolve the issue, scale up the replica counts of the operators.
-
Remove the modified
common-service
OperandRegistry.oc delete operandregistry common-service -n <namespace-where-foundational services-are-installed>
-
Scale up the
IBM Cloud Pak foundational services
andoperand-deployment-lifecycle-manager
operators.oc scale deployment operand-deployment-lifecycle-manager -n <namespace-where-foundational services-are-installed> --replicas=1
oc scale deployment ibm-common-service-operator -n <namespace-where-foundational services-are-installed> --replicas=1
-
Get the
common-service-operator
pod name.oc get pod -n <namespace-where-foundational services-are-installed> | grep common-service-operator
-
Delete the
common-service-operator
pod so that a new pod is created.oc delete pod <common-service-operator-pod-name> -n <namespace-where-foundational services-are-installed>