Skip to main content
  • EN

Background:

We have a multi-cluster setup with Openshift installed on On-Prem, AWS and Azure. On-Prem is designated as Primary and AWS, Azure clusters are discovered as Secondary in the Kasten K10 Multicluster UI. A S3 bucket is used for global location profile. We have created global snapshot and restore policy. We have annotated K8 supported CSI driver for Kasten volumesnapshot. 

We can take snapshot (on an application) from On-Prem cluster and restore it to AWS cluster using the same CSI driver (persistent volume created on same storage). Thus our goal achieved.

Concern:

We see from our storage console, that Kasten restore operation created a clone volume. In earlier version of Kasten (v5.5), we saw that this clone volume was mounted as persistent volume to the restored application. Now in v6.0.3, we see that clone volume was there but not mounted. We see a new volume is created and mounted as persistent volume to the restored application. However all the data from primary cluster is available on the new volume (restored application). 

We would like to know:

  1. why Kasten created a new volume but not using(mounting) the clone volume (for restored application) as in earlier version (v5.5). 
  2. Whether the application data was there in S3 bucket and copied to the new volume from the bucket during restoration. 

Thanks in Advanced. 

 

@jaiganeshjk 

@Shubhadip Pal Thanks for posting this question.

I see that you are trying to understand how k10 restores PVC. When does it use clones and when it creates a new volume to restore the data. 

I am not sure if this will give you the answer that you are looking for.

Below are two workflows that can happen during restore.

Restore from the snapshot  - Restore from local restorepoints involving storage snapshots uses cloning workflow to clone the snapshot into a new volume and use it for the application.

Restore from exported restorepoint - Restore from exported restorepoints involves creation of a new PVC, mount it into a restore-data pod and restore the data into it. Then we scale up the application pods to mount the same volume. 

 

PS: There is also export workflow that uses clone function to create a temporary volume that gets mounted to the copy-vol-data pods to facilitate exports. 

Comment


Terms & Conditions