Multi-node setup: Do they need same specs?

First question: I have been running ES on a single node (mostly because of budgetary constraints). Right now it has reached its full capacity and I am looking into adding nodes to it. However, again, because of budgetary constraints, I would like to know if it is possible to put it on a server that does not have the same specs as the current one.

Docs seem to imply that I should have 3 nodes — so for example, is it possible to keep my current node running on a 6-core machine and then having the additional two running on 4-core machines? The main reason I am asking is because the 6-core machine in fact costs twice as much as the 4-core and I would like to limit my cost if possible.

Second question: If I run snapshots, on Node 1, the data from Node 2 and 3 will be saved also, correct? When I used the S3 plugin to save, the performance is often unideal, and I find myself first snapshotting using the FS method and then sync’ing the locally backed up snapshots to S3 after it has been saved locally. And I wonder if I have a multi-node setup, that I would still be able to do this.

Yes. Please note that Elasticsearch will work out its thread pools given the underlying hardware specifications.
ref: https://www.elastic.co/guide/en/elasticsearch/reference/6.2/modules-threadpool.html

Correct. You can store snapshots of individual indices or an entire cluster
ref: https://www.elastic.co/guide/en/elasticsearch/reference/6.2/modules-snapshots.html

Thanks very much this is very helpful!

1 Like

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