Maintain PVs Autoscaling

This topic describes how to perform the following maintenance operations:

• Delete workloads

• Upgrade components

• Uninstall Zesty Disk for K8s

Delete workloads

The following rules apply to deleting:

1. Deleting the workload is the same as native K8s components:

   • If configured with WhenDeleted: Delete, the connected resources will be deleted.

   • If not configured that way, no further resources are deleted.

2. The Zesty PVC deletion process is as follows:

   • If the value of ReclaimPolicy is delete, the dummy PVC is deleted, and this triggers the deletion of the Filesystem and related PVCs and PVs.

   • If the value of ReclaimPolicy is retain, no further deletions occur, and the dummy PV (and its corresponding resources) stays intact.

Upgrade components

If you ever upgrade your cluster, you must be sure to upgrade PVs Autoscaling afterwards.

To upgrade components:

• Run these commands:
  

  helm repo update
  helm upgrade zesty-pvc zestyrepo/zesty

Uninstall PVs Autoscaling

You can perform a Helm uninstall only if there are no Filesystems of the Zesty storage class in the cluster. This is because uninstalling PVs Autoscaling components causes the native Pod migration of K8s to stop working,

To uninstall PVs Autoscaling:

• Run the following command:
  helm uninstall/delete zesty-pvc -name=<namespace>

The -name=<namespace> - argument is optional, needed only if PVs Autoscaling is installed on a dedicated namespace.

During the uninstallation, Zesty runs a Pod named “zesty-pvc-assert-uninstall-pre-delete-job” which runs pre-uninstallation validations. If the validations failed, this Pod will stay alive so its logs can be reviewed for further analysis.