High Ram.Percent

Hi ALL,

I have a 5 node ES 2.3.2 cluster all running as master-data enabled nodes. When I execute _cat/nodes i see all my nodes have ram.percent having > 95%.

Each of these machines have 28GB Ram and have allocated 14G as heap size for ES and all these are dedicated machines for ES( no other major applications/process are run on this machine)

Question:
What might be the cause for high Ram percent? What would be the consequences of high Ram utilization and what steps can I take to further to rectify this issue.

ES 2.3.2

host ip heap.percent ram.percent load node.role master name
...174 ...174 46 98 1.00 d m nodezy
...176 ...176 61 98 1.02 d m nodeaa
...173 ...173 47 98 1.26 d * nodezx
...175 ...175 61 98 1.30 d m nodezz
...172 ...172 50 92 1.20 d m nodezw
...170 ...170 43 98 1.73 d m nodezu
...171 ...171 65 97 0.62 d m nodezv

Elasticsearch relies on the heap as well as operating system file cache for optimal performance. The memory on the host is there to be used, and the operating system file cache will naturally fill up over time. If this memory however is needed for something else by the operating system, memory will be freed up by evicting data from the cache, so this should not be a problem.

Thanks for the explanation. Just Like GC's which frees up ES heap will file system cache be cleared automatically by ES say on some specific conditions or intervals? or will it remain there forever till a new process requests for memory

The operating system will release it when it needs to be used for something else, e.g. starting up a new process, so it should not cause any problems.

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