Undeployment fails or takes too long

If some Kubernetes resources from the Kaapana deployment are not in a Running or Completed state, undeployment might hang or fail.

Use Deployment Script Flags

Kaapana provides deployment script flags to help resolve common undeployment issues:

--no-hooks Purges all Kubernetes deployments, jobs, and Helm charts without running pre/post delete hooks. Use this if undeployment fails or runs indefinitely.
--nuke-pods Force-deletes all pods within Kaapana-related namespaces. Useful if stuck pods are blocking removal.

Run the deployment script with one or both of these flags:

./kaapanactl.sh deploy --no-hooks
./kaapanactl.sh deploy --nuke-pods

Manual Undeployment (Advanced)

If the above options don’t resolve the issue, you can manually undeploy Kaapana. > ⚠️ Warning: Only perform these steps if you are familiar with Kubernetes and Helm.

Step 1 – Remove Helm Releases

Check for existing releases:

helm ls -n admin
helm ls -n admin --uninstalling
helm ls -n default
helm ls -n default --uninstalling

Uninstall any Kaapana-related charts (e.g., platform-chart, admin-chart, project-charts):

helm uninstall -n <namespace> <chart-name> --no-hooks

Ensure that no Kaapana Helm releases remain.

Step 2 – Delete Namespaces

List current namespaces:

kubectl get namespaces

Remove any Kaapana-specific namespaces, such as:

services
admin
project-<name>

Delete each one manually:

kubectl delete namespace <namespace>

Step 3 – Clean Up Persistent Volumes

Check for remaining persistent volumes:

kubectl get pv

If needed, delete all persistent volumes:

kubectl delete pv --all