Cluster relocation

[sentaycan]

Hello, I am copying my PVC to a cluster in a different region as below, but now I want to delete my source cluster and use my destination cluster as source cluster, what is the most optimal way to do this?

kubectl --context west2 -n mysql create -f - <<EOF apiVersion: volsync.backube/v1alpha1 kind: ReplicationDestination metadata: name: mysql spec: rsyncTLS: copyMethod: Snapshot capacity: 8Gi accessModes: ["ReadWriteOnce"] storageClassName: gp3-csi volumeSnapshotClassName: csi-aws-vsc keySecret: east-west serviceType: LoadBalancer EOF
kubectl --context east2 -n mysql create -f - <<EOF apiVersion: volsync.backube/v1alpha1 kind: ReplicationSource metadata: name: mysql spec: sourcePVC: data-mysql-0 trigger: schedule: "*/5 * * * *" rsyncTLS: keySecret: east-west address: abad73aa2ca4441ed8c9e13f1095c453-95258c1d3ff95327.elb.us-west-2.amazonaws.com copyMethod: Snapshot volumeSnapshotClassName: csi-aws-vsc EOF

[tesshuflower]

@sentaycan
Once you have a PVC on the destination and you're happy with it (i.e. you've created a PVC from the snapshot created by the replicationdestination and it contains your data), then you will just have to do the reverse to setup a replication going in the other way.

That is, you'll need to create a replicationdestination on the old source (new destination) cluster, and your new PVC on the destination (that you now want to be the source) should be the sourcePVC for a new replicationsource.

[tesshuflower]

One note - on the new destination cluster, since it contains the old Source PVC, you can use this as your destinationPVC in your replicationdestination if you wish - this means the initial sync back from new source to new destination should be quicker as the destination PVC already contains a lot of the data.

[sentaycan]

Thank you, but I want to connect the PVC in DestinationCluster to my application, how can I do this, I couldn't do it :(

[sentaycan]

And my question was actually this.
I want to delete my source cluster and use the destination cluster as the main cluster.
What is the most optimal way to do this?