Hi Dika,

The export is basically your backup in simple terms. Yes there is a local snapshot created with csi (at least most of the time) but that is not really considered a backup. So your local snapshots get exported to a location (location profile) that can be S3 or NFS. 

You can recover from that location back to your local cluster on top of recovering from local snapshots.

If you want to recover the exported snapshots to another location (another Kasten deployment) make sure to get the import details from that job. https://docs.kasten.io/latest/usage/restore.html?highlight=restore

Disaster Recovery is when you have turned on that feature in Kasten. That essentially will create an export of all your Kasten settings to a location (location profile). In case of a disaster you would then be able to import all of your Kasten settings, including all the individual restore details from jobs, into a new Kasten setup at a DR site or locally to a new kasten deployment. https://docs.kasten.io/latest/operating/dr.html

It is best to have the location of the exported snapshots offsite just in case.

The location profile is the target of your backup policies and would be the source of your imports, be it either for single applications or for a complete DR recovery.

Hope that helps.

cheers

@Geoff Burke it means that, if i want to recover my app from source cluster to target cluster, both cluster must use same location profile.

source cluster will export and target cluster will import the snaphost.

is it right ?

Yes. You export to a location profile and then from the other Kasten deployment you connect to that location profile (providing the information required for each job or Kasten DR information, what I sent in the link above) and then import and restore.  In fact that is one way to have an automatic DR setup. There are posts about that here and I believe in the Kubernetes Korner as well.  

Hello, So If I have import details, I can create import policy on target cluster to restore my application instead of using DR.

In addition to Geoff's points, when performing the DR restore on the target cluster, ensure that the CRDs for your apps from the source cluster are also available on the target cluster. https://docs.kasten.io/latest/usage/migration.html#non-application-resources

@michaelxue Thanks ! My applications use openshift resources without any CRD.