Skip to main content

​
Preview the plan

The uninstall script is published alongside the installer, so it can be run without a repository checkout. It supports dry-run mode:
For a specific context and namespace:

​
Before you uninstall

The release owns the KubeBlocks definitions that every deployment references. Once they are gone, KubeBlocks can no longer reconcile or heal a deployment, so the script refuses to run while FalkorDB deployments exist. Delete them first, after copying any data you need to object storage. Pass --force only to uninstall anyway; the deployments keep running, but nothing manages them. The script keeps:
  • The backup repository (BackupRepo) and the backups in it. Reinstalling under the same release name and namespace adopts it again. Delete it with kubectl delete backuprepos.dataprotection.kubeblocks.io <name> when you no longer need the backups.
  • The bootstrap admin user Secret and the <release>-initial-admin-password Secret, unless --delete-namespaces deletes the namespace they are in.
When the installer made its bundled csi-hostpath-sc the default StorageClass, the script makes the original default StorageClass the default again.

​
Uninstall Enterprise only

This removes the FalkorDB Enterprise Helm release and retained FalkorDB addon resources. KubeBlocks is kept.
For automation:

​
Also uninstall KubeBlocks

​
Delete namespaces

If --uninstall-kubeblocks is not provided, --delete-namespaces deletes only the Enterprise namespace.

​
Delete CRDs

Deleting CRDs removes KubeBlocks and VolumeSnapshot API definitions from the cluster. This can affect other workloads that use those APIs.

​
Verify removal

If KubeBlocks was removed:
If namespaces were deleted: