I've just setup an ElasticSearch index and it's performing very poorly.
There's no information in the guide about running ElasticSearch from an
operational perspective, which is a frustrating experience as a newcomer.
Is there anything like MongoDB's production noteshttp://docs.mongodb.org/manual/administration/production-notes/ available
anywhere on the web? Are there rules of thumb about how to configure
ElasticSearch or troubleshoot it? I installed the .deb package, inserted
50M records, and the default out-of-the-box performance is very bad. The
data I'm working with is professional resumes and I'm trying to do searches
for job candidates with the Java client library.
Out-of-the-box, elasticsearch performs pretty well as far as default server
settings go. Set your JVM to use 50% of the available memory on the server.
Can you share your index mapping and an example query? Are you using facets
or nested/child objects? There are many ways to achieve poor performance
via inefficiencies in the mapping.
I've just setup an Elasticsearch index and it's performing very poorly.
There's no information in the guide about running Elasticsearch from an
operational perspective, which is a frustrating experience as a newcomer.
Is there anything like MongoDB's production noteshttp://docs.mongodb.org/manual/administration/production-notes/ available
anywhere on the web? Are there rules of thumb about how to configure
Elasticsearch or troubleshoot it? I installed the .deb package, inserted
50M records, and the default out-of-the-box performance is very bad. The
data I'm working with is professional resumes and I'm trying to do searches
for job candidates with the Java client library.
Ivan, thanks so much for sharing! Setting the heap size correctly made the
query 10x faster! I'll read the rest of the docs you sent and see if I can
squeeze out any other improvements, but as long as performance is
consistent at this level for the other queries I'll want to do, then I
should be good to go. I submitted a pull request to better document this
setting https://github.com/elasticsearch/elasticsearch/pull/3460 in the
start script, so that others have an easier time getting started.
Thanks again!
-Ben
On Wed, Aug 7, 2013 at 8:35 PM, Ivan Brusic ivan@brusic.com wrote:
Out-of-the-box, elasticsearch performs pretty well as far as default
server settings go. Set your JVM to use 50% of the available memory on the
server. Can you share your index mapping and an example query? Are you
using facets or nested/child objects? There are many ways to achieve poor
performance via inefficiencies in the mapping.
I've just setup an Elasticsearch index and it's performing very poorly.
There's no information in the guide about running Elasticsearch from an
operational perspective, which is a frustrating experience as a newcomer.
Is there anything like MongoDB's production noteshttp://docs.mongodb.org/manual/administration/production-notes/ available
anywhere on the web? Are there rules of thumb about how to configure
Elasticsearch or troubleshoot it? I installed the .deb package, inserted
50M records, and the default out-of-the-box performance is very bad. The
data I'm working with is professional resumes and I'm trying to do searches
for job candidates with the Java client library.
Thanks,
Ben
--
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.
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.