Postupgrade task

After a successful upgrade of your full cloud or hybrid deployment, if you don't want to roll back to a previous version, remove remnants of the Elasticsearch solution.

Removing Elasticsearch

CAUTION:
If you remove remnants of the Elasticsearch solution, you cannot roll back to a previous version.
Complete the following steps to remove the nasm-elasticsearch and common-datarouting pods from your full cloud or hybrid 1.6.13 deployment.
  1. Wait until the noi or noihybrid custom resource (CR) for your upgrade moves to the reconciled phase. To ensure that version 1.6.13 is displayed for both version and reconciledVersion, run the oc get csv command.
  2. To ensure that all jobs during the upgrade are completed, run the oc get jobs command.
  3. Locate the StatefulSet (STS) for all instances of Elasticsearch.
    oc get sts -l app.kubernetes.io\/name=elasticsearch
    		NAME                   READY   AGE
    		ocm480-elasticsearch   1/1     82m
  4. Delete the STS for all instances of Elasticsearch.
    oc delete sts -l app.kubernetes.io\/name=elasticsearch
    		statefulset.apps "ocm480-elasticsearch" deleted
  5. Wait for all replicas of Elasticsearch STS to terminate. The following command watches for status transitions of the pods. After all pods transition, you can terminate the command.
    oc get po -w -l app.kubernetes.io\/name=elasticsearch
  6. Delete the Elasticsearch curator cronjob.
    oc delete cronjob -l app.kubernetes.io\/name=ibm-common-elastic-curator
    		cronjob.batch "ocm480-curator-pattern-metrics" deleted
  7. Delete the common-datarouting deployment for all instances of common-datarouting.
    oc delete deploy -l app.kubernetes.io\/name=common-datarouting
    		deployment.apps "ocm480-common-datarouting" deleted
  8. Depending on the persistent volume claim (PVC) policy, the Elasticsearch PVC might need to be manually removed. However, the Elasticsearch PVC is the final remnants of all data in Elasticsearch. After you delete the PVC, you cannot roll back to previous versions.
    oc get pvc -l app.kubernetes.io\/name=elasticsearch
    		NAME                          STATUS   VOLUME                                     CAPACITY   ACCESS MODES   STORAGECLASS      AGE
    		data-ocm480-elasticsearch-0   Bound    pvc-4b08bbe4-6a9e-4016-b279-dfa395f98760   75Gi       RWO            rook-ceph-block   89m
    
    	oc delete pvc -l app.kubernetes.io\/name=elasticsearch
    		persistentvolumeclaim "data-ocm480-elasticsearch-0" deleted

Post-rolling back

If you upgraded to version 1.6.13 and did not remove remnants of the Elasticsearch solution, you can roll back to the previous version. After you roll back to the previous version, decommission the workloads that were deployed with version 1.6.13. If the workloads are decommissioned, when you upgrade to version 1.6.13 after a roll back, the Elasticsearch data is migrated again. Before you attempt another upgrade post-rolling back, complete the following steps.
  1. Delete any AIDL (ea-datalayer) deployments.
    oc delete deploy  -l app.kubernetes.io/\component=aidl
  2. Delete the AIDL curator cronjob.
    oc delete cronjob -l app.kubernetes.io/\component=aidl-curator
  3. Delete the Elasticsearch migration job so that it can be reattempted when you next upgrade.
    oc delete job     -l app.kubernetes.io/\component=elasticmigration