Heap Usage is not balanced on all the nodes


#1

hi All,

I have a cluster contains 15 data nodes, the heap usage on every nodes are having huge difference. However all our data nodes are load balanced from indexer side. the highest one can be more than 80% and lowest one is 20%
is this normal? if not, how can I troubleshoot this issue?

all the data nodes configuration is same. (heap is set at 31gb and mlock = true)

versions we are using:
logstash 2.4.1
elasticsearch 2.4.1
kibana 4.6.1


(Zachary Tong) #2

It's not necessarily un-normal :slight_smile:

Heap usage is largely driven by utilization by shards that are local to that machine. So you may have some shards co-located on nodes that are active and using a large amount of memory, while other nodes have shards that are small / don't use much heap / not very active.

You could look through the Node Stats to identify where heap is going on the nodes with high usage.


(system) #3

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.