Hello Folks,
I posted on this topic yesterday and didnt hear back, thought I'd try
again, has anyone used the elasticsearch java jest library and done perf
tests with it running under tomcat. We have a rest service that abstracts
away the jest library and performs crud functionality against an
elasticsearch cluster and we're currently having some issues with running
out of operating system sockets/file descriptors during our load test when
using this library within a very short period of time, before I make any
major changes I plan to:
Tweak the associated httpclient connection manager configs
Tweak environment linux settings
Finally if the above don't make a difference I plan to:
3) Rewrite the code using the elasticsearch java API instead
Would love to hear from the community about any learning around perf tests
around the library.
I have replied to you through our support but maybe it is not reached.
Well first please create only one JestClient it is singleton and configure
it as multi threaded. That is very much possibility why you are having too
many open sockets errors.
To see all in action check JestClientFactoryTest.multiThreadedClientCreation
for sample usages.
On Tuesday, April 9, 2013 6:24:22 PM UTC+3, Saikat Kanjilal wrote:
Hello Folks,
I posted on this topic yesterday and didnt hear back, thought I'd try
again, has anyone used the elasticsearch java jest library and done perf
tests with it running under tomcat. We have a rest service that abstracts
away the jest library and performs crud functionality against an
elasticsearch cluster and we're currently having some issues with running
out of operating system sockets/file descriptors during our load test when
using this library within a very short period of time, before I make any
major changes I plan to:
Tweak the associated httpclient connection manager configs
Tweak environment linux settings
Finally if the above don't make a difference I plan to:
3) Rewrite the code using the elasticsearch java API instead
Would love to hear from the community about any learning around perf tests
around the library.
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.