Can I take backup of indices from one cluster and restore it to another cluster

Can I take backup of indices from one cluster and restore it to another cluster by simply copying the snapshot/backup repository folder and sending it to another cluster and from there I will perform restore operation is it possible.

Yes. That should work.

Yes, as long as the version of Elasticsearch on the destination cluster is the same or newer than the source cluster.

I tried it but it did not work. I copied the indices folder that was one one system and simply moved them to one location on different system and from there I tried the restore operation. Did I miss anything here.

Which version of Elasticsearch are you using in the two clusters?

What is the size of the two clusters?

How are you taking the snapshots? Shared filesystem?

Exactly how did you go about moving the repository?

How did you try to restore?

Q] Which version of Elasticsearch are you using in the two clusters?
A] On the system where I have taken snapshot I am running 8.7.0 and on the system where I have copied the folder containing indices runs 8.3.2.

Q] What is the size of the two clusters?
A] Approximately 15GB

Q]How are you taking the snapshots? Shared filesystem?
A] I took snapshot by first creating a repository and registering that repository, after that I created a snapshot using _snapshot operation and took snapshot of all indices in a snapshot. And no I have not created a share filesystem.

Q] Exactly how did you go about moving the repository?
A] I am using traditional approach copy and paste. Copy the folder containing snapshot pasting them in another system and performing restore operation through kibana dev tools.

Q] How did you try to restore?
A] On a different machine I created a repo regsistered it and inside that repo I pasted the copied folder and on kibana dev tools I tried _restore operation.( I know this is not how it is done ). Please tell me a workable solution.

A snapshot taken from a 8.7.0 cluster can only be restored into a cluster running Elasticsearch 8.7.0 or later.

How many nodes do the clusters have?

target cluster has 1 node and source cluster also has one node

Is it necessary to have all the cluster on same network.

Both nodes need to have access to the snapshot repository. If you have single node clusters you should be able to copy the data over and mount the repositoty as read-only.

1 Like

can you direct me to any documentation regarding it. (how to configure a repo as read-only)

1 Like

Is it necessary to have all the cluster on the same network.

No, not if you copy the snapshot repository over and restore locally.

1 Like

Ok, Thank you very much

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.