Master node heap cost

Hello,I have a problem about my cluster's master node heap cost!

after some days ,the heap of master node cost chart is this:

after i restart master node,heap used chart is this:

every day ,i have job to remove ,close,force merge indices(daily indices),so es data size is stable,but the heap used bigger and bigger , i have to adjust heap size or restart master node(dedicated master node)

my es version is 5.0.1, i guess is there a memory leak problem?

search for a help!

thanks in advance!

Do you have any plugins installed? Do you see any increase in the number of pending task on this master?

x-pack,monitor plugin enabled.

there is no pending task,the load of my service is low.

after indices manage(delete,close,force merge),is it could reduce heap cost of master node?

how to optimize heap cost of master node?

thanks for you reply!

Hmm, could you create a heap dump of your master node when it runs out of memory next time to see how the heap memory is used? You can analyze it yourself or email it to me, just don't post it here since it might contain sensitive information about your cluster setup.


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