Helm uninstall leaves chi pod, sts, svc around

[tiina303]

Bug description

Helm uninstall leaves chi around

Expected behavior

Everything is deleted

Actual behavior

Clickhouse pod, sts, svc don't get deleted.

$ helm uninstall posthog
$ kubectl get all
NAME                            READY   STATUS    RESTARTS   AGE
pod/chi-posthog-posthog-0-0-0   1/1     Running   0          19m

NAME                              TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)                         AGE
service/chi-posthog-posthog-0-0   ClusterIP   None             <none>        8123/TCP,9000/TCP,9009/TCP      18m
service/clickhouse-posthog        NodePort    10.245.199.230   <none>        8123:31175/TCP,9000:32642/TCP   19m

NAME                                       READY   AGE
statefulset.apps/chi-posthog-posthog-0-0   1/1     19m

How to reproduce

1. install posthog
2. helm uninstall posthog -n posthog
3. kubectl get all -n posthog

Environment

• Deployment platform (DigitalOcean): please provide
• k8s version (server version from kubectl version): 1.21
• Chart version/commit (helm list -n posthog): master
• Posthog version: master

Additional context

Related problem after manually deleting the chi stuff running helm install doesn't create chi stuff, which can be fixed by running a helm upgrade , but helm install should work.

[macobo]

This is caused by finalizers in the clickhouse-operator.

It might be a 'feature' in that it doesn't drop the data.

A one-command workaround is kubectl patch chi posthog -n posthog -p '{"metadata":{"finalizers":null}}' --type=merge

Also see how we work around this in the kubetest test suite.