activemq-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From chrisr <chris.ridm...@gmail.com>
Subject Clustered LevelDB/Zookeeper AMQ producer response times spike during a failover
Date Tue, 13 Oct 2015 16:16:25 GMT
Hello,

I am using ActiveMQ v5.12.0 using a clustered leveldb/zookeeper setup.

I am doing some performance testing of the setup during failover scenarios. 
We noticed that during a failover our JMS producer latency jumps up from an
average of 200-400ms to 25000ms.  This spike lasts for 2 minutes.

We are not using connection pooling.  Unfortunately I can't give you a test
case to reproduce at this time, but am wondering if you guys have
experienced this in your own benchmarks or performance runs.

I went through the docs and didn't find anything immediately obvious that I
needed to tune.  Could you guys make a recommendation on things that I can
try to try to reduce the spike duration or intensity?

Thanks!
-Chris





--
View this message in context: http://activemq.2283324.n4.nabble.com/Clustered-LevelDB-Zookeeper-AMQ-producer-response-times-spike-during-a-failover-tp4702912.html
Sent from the ActiveMQ - User mailing list archive at Nabble.com.

Mime
View raw message