Originally from the User Slack
@Igor_Domrev: Hello, I have a question.
We are currently running scylla in production on k8s , on spot instances. Using scylla k8s operator.
It works like a clock for three years already. 
Our problem is that we accidentally created an un-balanced deployment of scylla on different availability zones.
1 node in the first AZ
2 nodes in the second AZ
We want to fix that and move to a balanced deployment.
However, because each pod has a PVC in its zone.
We first need to create a new PVC in the new zone.
we are currently on scylla 5.2 community. cluster of three nodes.
What is the best strategy todo proceed ?
If we just delete the disk to one of scylla pods, when it re-spawns and provisions a new and empty disk in the new zone.
Will it just re-sync all the data from the other cluster nodes ?
We can also create PVC in the new ZONE from backup and make the new pod to mount it.
Will appreciate any advise on the matter.
Thanks in advance
@Maciej_Zimnoch: Use replace dead node procedure: https://operator.docs.scylladb.com/stable/resources/scyllaclusters/nodeoperations/replace-node.html
Replacing a Scylla node | ScyllaDB Docs
@Igor_Domrev: TY for the reply.
will this work with operator 1.8 ?
can i upgrade the operator to 1.17 without upgrading the scylla cluster ?
@Maciej_Zimnoch: I don’t know, 1.8 is EOL. 5.2 is also EOL and is not supported by 1.17