I will try and reproduce your problem on 5.3.1 build, but in the mean time, can you make sure your ElasticSearch instance is not under heavy load or excessive garbage collection.
In that case , you probably need more resource for you ElasticSearch installation, you can turn up the timeout in the config file, but that will only help so much.
My 2 data nodes have 16GB physical memory.
Looks like somebody had adjusted the JVM heap size to be 16GB as well!
I have set this to 12GB and things are behaving much better now.... no crashes.
My ES nodes are are on a vm host that is very low on memory. Will be rectifying that shortly.
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.