Here is something worth sharing, I think. For Logsene http://sematext.com/logsene/, We've been trying out EC2 R3 instances
(specifically r3.large) with Elasticsearch (1.1.x and 1.3.x including
1.3.2) with the latest Java 7 update 67. We tried this on Ubuntu 14.04
AMI, as well as Amazon Linux AMI.
In all cases we see Elasticsearch (or the JVM?) periodically "freezing" for
minutes at a time.
Here is a graph that shows
this: https://apps.sematext.com/spm-reports/s/kOn7sFUCe2 -- uncheck all
checkboxes except the one on the right -- "total opened HTTP conns
(total_opened)'" and you will see the pauses on 3-4 servers happening.
We'll move off of R3s, but wanted to share this in case anyone else runs
into the same problem. We don't have this problem on instances of other
types we are running.
We are actively using the r3 series in EC2 and have not experienced any
issues. We are also using Java 7u67 but we are using Ubuntu 12.04
exclusively. We have tested with ES 1.1.1 and 1.3.2 and have not run into
any issues. We are predominantly using the r3.2xls but I think we have used
some r3.xls and r3.larges. Are you doing anything special with networking?
Drivers? EBS optimized?
We are actively using the r3 series in EC2 and have not experienced any
issues. We are also using Java 7u67 but we are using Ubuntu 12.04
exclusively. We have tested with ES 1.1.1 and 1.3.2 and have not run into
any issues. We are predominantly using the r3.2xls but I think we have used
some r3.xls and r3.larges. Are you doing anything special with networking?
Drivers? EBS optimized?
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.