How to delete a cluster and its current state?

General
#1

I have this cockroachdb cluster deployed with helm: helm install k8crdb --set Secure.Enabled=true cockroachdb/cockroachdb --namespace=thesis-crdb

helm list --namespace=thesis-crdb

NAME    NAMESPACE       REVISION        UPDATED                                 STATUS          CHART                  APP VERSION
k8crdb  thesis-crdb     1               2021-01-31 23:58:41.0618195 +0100 CET   deployed        cockroachdb-5.0.4      20.2.4

kubectl get all --namespace=thesis-crdb

NAME                                READY   STATUS      RESTARTS   AGE
pod/k8crdb-cockroachdb-0            1/1     Running     0          3m27s
pod/k8crdb-cockroachdb-1            1/1     Running     0          3m27s
pod/k8crdb-cockroachdb-2            1/1     Running     0          3m27s
pod/k8crdb-cockroachdb-init-kdzwf   0/1     Completed   0          3m27s

NAME                                TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)              AGE
service/k8crdb-cockroachdb          ClusterIP   None            <none>        26257/TCP,8080/TCP   3m27s
service/k8crdb-cockroachdb-public   ClusterIP   10.105.118.27   <none>        26257/TCP,8080/TCP   3m27s

NAME                                  READY   AGE
statefulset.apps/k8crdb-cockroachdb   3/3     3m27s

NAME                                COMPLETIONS   DURATION   AGE
job.batch/k8crdb-cockroachdb-init   1/1           33s        3m28s

Now I want to delete this cluster: helm uninstall k8crdb --namespace=thesis-crdb
And this works.

However, when I redeploy with the same command: helm install k8crdb --set Secure.Enabled=true cockroachdb/cockroachdb --namespace=thesis-crdb

The state has been saved from how it was before it got deleted. I wanna start a completely new cluster with the same name and where the old dead/commissioned nodes are gone. I even tried to delete the cockroachdb repo: helm repo remove cockroachdb
but it picks up from the previous state …

#2

Hi Andreas! It sounds like you need to remove the PersistentVolumeClaims and PersistentVolumes of your cluster.

If you run kubectl get pvc you should see a handful of claims with names that match the pod names from your post (k8crbd-cockroachdb-x). You can delete them with kubectl delete pvc <name>. If the reclaim policy is set to RETAIN, you’ll have to manually delete the persistent volumes in a similar fashion.