Cross cluster replication mentions that there is an option for HA (High Availability) that offers replication and automatic fail over.

A few questions about replication:

  1. Is this active replication?
  2. Is there any lag in replication?
  3. Does it guarantee no data loss in case of failover?

I am asking because I haven't seen replication as a feature offered in open source version of elasticsearch. I am curious how does this work internally. That would help us make right decision/s going forward. Thank you so much.

Well it depends on what you mean by DC? In the context of cloud a DC is a datacenter which is in the same region as the other DCs. Which means almost no network latency.

It's not using any specific code here. XDCR is coming in future versions of elasticsearch and it will allow having DCs across multiple regions.


What I am trying to understand is how does it guarantee HA. If one cluster falls over for some reason, the other cluster (probably in other datacenter in same region or different region) needs to persist same copy of data. This means either the data is ingested to both the clusters and both clusters are in sync or there is some background replication that ensures the data is in sync.

Please let me know if you need more clarification.

It's only one cluster in the context of cloud. Not multiple clusters.

A cluster has 3 nodes. Each node is in one DC, sharing the same region.

Ah, got it! thanks for clarifying.

BTW, what's XDCR? Cross DC Replication? Do you have any document that I can reference ?

Yes it is.

There is a small mention of it at

Thank you!

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