Clustering data consistency question


I have a cluster of 3 nodes where each node has all roles (master, data, ingest).
I use logstash which pushes the data to elasticsearch.

When is the ack sent to logstash? After the data has been replicated or just when it was saved to the ingest node?

Let's say I have node A,B,C. Logstash is sending a bulk index request to A. Is it possible that data is lost, if A crashes after sending the ack to logstash but before the data is replicated to B or C?

Thanks, Andreas

Elasticsearch only acknowledges a write request after all the documents in that request have been durably written to disk on all in-sync shard copies (primaries and replicas). See the reference manual for more details.

thanks for the fast reply.

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