Preference local on a transport client

Elastic Stack Elasticsearch
1

Hi,

I am using ES 1.3.2 and setup a cluster with 3 nodes, having about 350,000
docs right now. (1 shard, 2 replicas)

In my service I create a transport node connecting to the cluster, and use
Java API to search.

Then I benchmarked my service - each HTTP request to my service will
trigger about 12 to 15 queries (sent with multi-search api).

  • Without preference set to local, all requests get slow when
    concurrency goes high.
  • With preference set to local, some of the requests are as fast as low
    concurrency under high concurrency, but some of them are super slow

So I want to know how does this config affect the performance in my setup
(1 transport --> 3 data nodes).
Also I don't understand why some node can handle the queries very fast
while the others are super slow... I thought in my setup every node should
have complete data stored.

Thanks :slight_smile:

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elasticsearch+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/elasticsearch/a1a233a5-9988-4ba5-9c5c-14d0b3fc3a2e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.