How to reuse PVCs and PVs after helm uninstall/install?

[zumbuschms]

We have a RabbitMQ cluster with three nodes deployed to an Azure Kubernetes Services cluster. When we run helm uninstall for the charts including the RabbitMQ cluster deployment, the physical volume claims for the nodes are deleted leaving the phyical volumes behind in Released state. We have the PVs configured to use a custom storage class with the reclaim policy set to Retain. As these PVs still carry the claim reference to those deleted PVCs, they are not reconnected to new PVCs when I reinstall the helm charts to redeploy the RabbitMQ cluster. With this behavior we lose the data that was accessed by the cluster.

Is there an option to keep the PVCs and reconnect them to the newly created pods after the reinstallation?

[zumbuschms]

I fiddled around a bit more and tried setting the persistentVolumeClaimRetentionPolicy via either the StatefulSet spec or the override spec. Neither worked. I stumbled over owner references that are defined for the PVCs with the RabbitMQCluster as their owner and blockOwnerDeletion set to false. When uninstalling the cluster, Kubernetes automatically deletes dependant resources and seems to ignore the persistentVolumeClaimRetentionPolicy.