How to create Elasticsearch Dedicated Replica Cluster

Hello Everyone:

I am using Elasticsearch v7.8.0 cluster having indices in 100TB+.

I am looking for a backup cluster for my running PROD cluster , a cluster where my DATA would reside in READ-ONLY Manner while it still gets data SYNC from READ-WRITE Cluster.

Is there any provision where I can have dedicated ES Cluster as READ-ONLY of another ES Cluster which is in READ-WRITE ?

Thank you for your consideration and responses.

This sounds pretty much what Cross Cluster Replication is for?

Hello @spinscale

Thanks for your response.

I was away for some other work ... hence could not respond ...

Thanks for this suggestion .. this could be my solution too .. I am going to read its documentations.

If any doubts , I shall come back on this.

Thank You.

This CCR , is Platinum level feature. We are having BASIC license.

Any option , that would help us under current license.

2.scroll reindex;

Yes , Snapshots we have , but it takes time [ for backup + restoration ] as our cluster is 100TB+ in size.

I would explore scroll reindex option, now.

Thank You @casterQ

Shanshot backup and recovery are incremental,maybe you can do something on an incremental basis

Yes , they are by default incremental way only. We are taking them like that only ..

Incremental our side takes 2 hours of time...

In regards to this , could you please tell me is there way by which i can take every time full backup , rather then incremental ?

SCROLL REINDEX = means specifying SIZE option during reindex in source ...

  1. This is limited to 10k only..
  2. That would require everything reindexing and checking whole index again and again in time interval..

I am looking for some option which would be at run-time SYNC to another replica .. to have immediate availability of data , in case of primary fails ..

Alternative to CCR in basic license ..

If i do not get, than will have to take up one of the available options .. Hence i am exploring if there are any more options via this forum.

i think you can create a new repository and create the snapshot in it, it's will be a full backup

maybe you can realize double write by yourself

Okay ...

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