Symptoms Prevention, Resolution or Workaround
Persistent volumes don’t get created on the target cluster. Run kubectl describe on one of the pods of replication controller and see if event says Config update won't be applied because of invalid configmap/secrets. Please fix the invalid configuration. If it does, then ensure you correctly populated replication ConfigMap. You can check the current status by running kubectl describe cm -n dell-replication-controller dell-replication-controller-config. If ConfigMap is empty, please edit it yourself or use repctl cluster inject command.
Persistent volumes don’t get created on the target cluster. You don’t see any events on the replication-controller pod. Check logs of replication controller by running kubectl logs -n dell-replication-controller dell-replication-controller-manager-<generated-symbols>. If you see clusterId - <clusterID> not found errors then be sure to check if you specified the same clusterIDs in both your ConfigMap and replication enabled StorageClass.
You apply replication action by manually editing ReplicationGroup resource field spec.action and don’t see any change of ReplicationGroup state after a while. Check events of the replication-controller pod, if it says Cannot proceed with action <your-action>. [unsupported action] then check spelling of your action and consult the Replication Actions page. Alternatively, you can use repctl instead of manually editing ReplicationGroup resources.
You execute failover action using repctl failover command and see failover: error executing failover to source site. This means you tried to failover to a cluster that is already marked source. If you still want to execute failover for RG, just choose another cluster.
You’ve created PersistentVolumeClaim using replication enabled StorageClass but don’t see any RGs created in the source cluster. Check annotations of created PersistentVolumeClaim. If it doesn’t have annotations that start with then please wait for a couple of minutes for them to be added and RG to be created.
When installing common replication controller using helm you see an error that states invalid ownership metadata and missing key "": must be set to "Helm" This means that you haven’t fully deleted the previous release, you can fix it by either deleting entire manifest by using kubectl delete -f deploy/controller.yaml or manually deleting conflicting resources (ClusterRoles, ClusterRoleBinding, etc.)
PV and/or PVCs are not being created at the source/target cluster. If you check the controller’s logs you can see no such host errors Make sure cluster-1’s API is pingable from cluster-2 and vice versa. If one of your clusters is OpenShift located in a private network and needs records in /etc/hosts, exec into controller pod and modify /etc/hosts manually.
After upgrading to Replication v1.4.0, if kubectl get rg returns an error Unable to list ", Resource=dellcsireplicationgroups" This means kubectl still doesn’t recognize the new version of CRD after upgrade. Running the command kubectl get<rg-id> -o yaml will resolve the issue.