I have a 5node cluster all data, client and master and node 1,2,3 are on 6.3.1 version and node 4 and 5 are on 6.5.1 version. All nodes are 16cpu, 64gb ram, 32 gb heap and 500gb disk.
Load and cpu on nodes 1,2 and 3 are very high and node 4 and 5 is pretty low.
Latency is increased on 1,2 and 3 nodes.
Can someone help me fix this such that load is distributed among all nodes equally.
The first thing to try is to complete your upgrade so that all the nodes are on the same version, particularly since the problems are correlated with the versions of the nodes.
Also note that v6.5.1 has a known issue, fixed in v6.5.3. I recommend avoiding that version, and ideally you should upgrade everything to the latest in the series (v6.7.1 at time of writing).
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.