Restoring Watson Studio

The following cleanup procedures are required before restoring Watson Studio: deleting install jobs and deleting artifacts related to any active environments or running jobs. The procedures are applicable to both Portworx and NFS.

Deleting install jobs

These Kubernetes jobs are created during the installation of Watson Studio, Watson Machine Learning or Watson Knowledge Center. They might have already been deleted, but if they weren’t, delete them now. The job names are:

You can copy the following code sample into a script file to delete these jobs:

#!/bin/bash
jobsToDelete=("dsx-requisite-pre-install-job" "ibm-environments-prod-ccs-init-job" "ibm-notebooks-prod-wsl-init-job" "change-pvc-permissions" "dsx-userhome-pre-install-job")
for job in ${jobsToDelete[@]}; do
  oc delete job $job
done

Note: This step is not necessary if you’re restoring to a freshly installed Cloud Pak for Data.

When you’re restoring to a Cloud Pak for Data that has previously been used, this step will stop the active environments and running jobs that were associated with the previous state of the cluster.

These artifacts include deployments, services, jobs and cron jobs. For each artifact, there is an associated  secret.  You can identify the artifacts by listing all secrets that contain the label created-by=spawner. Each of these secrets have a correlated artifact of one of the other types, with a corresponding name. For example, if there is a deployment named example1, there will be secret with that label with the name example1-sct.

Delete all these artifacts along with their associated secrets.