Rolling back Netcool Operations Insight on Red Hat OpenShift

Use these instructions to roll back from 1.6.5 of Netcool® Operations Insight® to a previously deployed 1.6.4, using the Red Hat® OpenShift® Operator Lifecycle Manager (OLM) user interface (UI), or the command line.

1. If Netcool Operations Insight 1.6.4 was upgraded to version 1.6.5 using airgap, then, before you rollback from version 1.6.5 to 1.6.4, you must set the image registry back to the 1.6.4 Docker image repository. Otherwise the cem-operator and asm-operator pods will fail with ImagePullError errors.
   1. Edit the noi-operator deployment, and find the key-value pair for OPERATOR_REPO. The value of this is set to the 1.6.5 airgap target registry. Replace this value with the Netcool Operations Insight 1.6.4 image registry where the Netcool Operations Insight 1.6.4 Passport Advantage® (PPA) package is uploaded, for example image-registry.openshift-image-registry.svc:5000/<namespace>.

      oc edit deploy noi-operator

   2. If the Netcool Operations Insight 1.6.4 image registry is authenticated and requires a pull secret, then edit the noi-operator serviceaccount and add this secret in the imagePullSecrets section.

      oc edit serviceaccount noi-operator

2. Rollback can be performed from the command line or from the OLM UI.
   To rollback from the command line, use: oc edit noi and change the version back to version 1.6.4.
   To rollback from the OLM UI, navigate to Operators > Installed Operators > NOI and then select the Cloud Deployment tab if your deployment is only on Red Hat OpenShift, or the Hybrid Deployment tab if you have a hybrid deployment that is on Red Hat OpenShift and on-premises. Select Edit NOI and then the YAML tab. Change the version back to version 1.6.4 and save the changes.
3. Delete the noi-topology-system-health-scheduledjob job and the noi-full-topology-system-health-cronjob cron job by running the commands:

   oc delete job noi-topology-system-health-scheduledjob

   oc delete cronjob noi-full-topology-system-health-cronjob

   Verify that the cron job is recreated by running oc get cronjob. The output should be:

   NAME                                      SCHEDULE        SUSPEND   ACTIVE   LAST SCHEDULE   AGE
   noi-full-curator-pattern-metrics          0 0 * * *       False     0        <none>          143m
   noi-full-healthcron                       1 * * * *       False     0        28m             143m
   noi-full-register-cnea-mgmt-artifact      1,*/5 * * * *   False     0        4m10s           143m
   noi-full-topology-system-health-cronjob   */5 * * * *     True      0        <none>          44m

4. Obtain the metrics deployment by running the command:

   oc get deployment | grep -i metric

   The output is the following:

   noi-metric-action-service-metricactionservice       0/0   0      0      34h
   noi-metric-api-service-metricapiservice          0/0   0      0      34h
   noi-metric-ingestion-service-metricingestionservice    0/0   0      0      34h
   noi-metric-trigger-service-metrictriggerservice      0/0   0      0      34h

   Delete the metrics deployments by using the command:

   oc delete deploy  <deployment name>

5. Delete the ibm-hdm-analytics-dev-v3-evt-pi-processor and common-datarouting deployments:

   oc delete deployment <release-name>-ibm-hdm-analytics-dev-v3-evt-pi-processor
   oc delete deployment <release-name>-common-datarouting