Hello team, since 6.4.0 I noticed that installing a plugin successfully makes Elasticsearch goes down. Then I must restart it manually and then all it's fine.
Is this an addition in 6.4.0? Why it happened?
Regards,
Jesús
Moved this to Elasticsearch forum as it is an ES question.
Installing an elasticsearch plugin should be done when elasticsearch is shut down. However, it should not cause the system to fail. Can you describe more concretely, without output from your commands, what you are seeing, including logs from elasticsearch when it stops?
Sorry but what I meant is a Kibana plugin. Once the Kibana plugin installation finishes, Elasticsearch goes down with no reason . After restart Elasticsearch all is just fine...
Why?
Regards!
Can you show your elasticsearch log around the time it goes down?
Ok @rjernst after some research I've found the cause from this situation. An out of memory error:
Aug 28 18:24:48 localhost.localdomain kernel: Out of memory: Kill process 1122 (java) score 255 or sacrifice child
Aug 28 18:24:48 localhost.localdomain kernel: Killed process 1122 (java) total-vm:4832964kB, anon-rss:459236kB, file-rss:0kB, shmem-rss:0kB
Aug 28 18:24:48 localhost.localdomain systemd[1]: elasticsearch.service: main process exited, code=killed, status=9/KILL
Aug 28 18:24:48 localhost.localdomain systemd[1]: Unit elasticsearch.service entered failed state.
Aug 28 18:24:48 localhost.localdomain systemd[1]: elasticsearch.service failed.
The optimizing procedure has a RAM peak and it causes Elasticsearch being killed. The machine was a bit overloaded so I understand it's not an error itself. Now my question is why is Elasticsearch the target to be killed instead other components or other installed software?
Regards
This reason would be platform dependent, but most likely elasticsearch is the highest user of memory on the system, so a prime target for the OOM killer when there is memory pressure.
This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.