I'm new to the forum.
I've upgraded my cluster from 1.2 to 1.7.3 version. Since then, we have lot of problems and our cluster is not very stable.
One of the critical problem is that we have lot of GC, and when looking at the _**_nodes/stats ** api, I can see that the memory usage is much higher than it used to be in 1.2 (around 60-80% while it was around the 40% in the past, going down and up periodically, with the time) .
The main symptom, from client perspective, is that we have an intermittent high latency during query (most of the time, latency is short but from time to time, it can be very high, up to one minute). From our investigation, it comes with a GC on one of our Elastic nodes...
As well, we started to have "latency/reject" during indexing .
Any one has experienced same issue with this Elastic 1.7.3 upgrade ?
We are using
- JDK 1.7.0_60
- swap mlockall