Circuit breaker Open for http_request

(Stephanus Tedy) #1

Cluster setup :
10 nodes , with roles:

  • 7 dedicated data only node (8 core, 16gb memory)
  • 1 data and master (8 core, 16gb memory)
  • 2 master, ingest and coordinator only node (4 core, 8gb memory)

Elasticsearch version 5.2.0

I use nginx in front of ES cluster, nginx will request to 2 dedicated coordinator node.

I have an issue with circuit breaker, this error occures every 1 week
[parent] Data too large, data for [<http_request>] would be larger than limit of [2982071500/2.7gb]

what is actually happens ? and how to solve this.
this problem happens, after I use this dedicated coordinator architecture,
which it helps a lot to reduce load on data node.


(Jason Tedor) #2

There were some bugs with accounting of releases from the circuit breaker fixed in 5.2.2, and some other relevant fixes in 5.6.3 and 5.6.4. I would encourage you to at least upgrade to 5.2.2, and if you can, all the way to 5.6.4.

(system) #3

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