Which network is used for shard replication?

(Yu Watanabe) #1


I am diagnosing the reason for bulk rejects which I am seeing in my environment.

I read below article and document understand the mechanism of replication.

  1. https://www.elastic.co/blog/why-am-i-seeing-bulk-rejections-in-my-elasticsearch-cluster

  2. https://www.elastic.co/guide/en/elasticsearch/guide/2.x/distrib-write.html

What I would like to ask is after write to primary succeeds how are write requests is sent to the other node? Does it use bulk request again , or does it use internal communication ( Transport protocol ) ?

(David Turner) #2

All node-to-node communication in a cluster uses the transport protocol (as opposed to HTTP, which is only used for client-to-node communication).

Perhaps a minor point, but note that a write to the primary doesn't succeed until after it's been replicated successfully to every replica.

Yes, they are bulk requests, in the sense that a single primary-to-replica request may contain many indexing operations.

(Yu Watanabe) #3

Thank you for the reply.

(system) #4

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