Reviving Elasticsearch 5.0.0

(Dimitris Filippou) #1

My Elastic Stack is correctly up and running with Logstash , Kibana and Elasticsearch. Elasticsearch is running as a daemon (/etc/init.d/elasticsearch start|stop etc...).

Sometimes due to low memory Elasticsearch fails and kibana status turns red. I obviously need to upgrade my memory and i did to 8 GB but still i want to be 100% sure that it will never crash/fail .

My question is how can i get Elasticsearch to run again as a daemon after it fails?

It fails rarely but it does.. so when it does i want it to run again.

My Elasticsearch version is 5.0.0 , im using ubuntu 16.04 and i followed these steps

in addition , supervisor is not ideal for this issue because it can't handle daemons.

(Mark Walkom) #2

Well you can setup something to autorestart ES, but that never solves the problem of why it stopped.
That's a whole other thing to deal with and I am not sure there is a simple solution.

(Alain Rastoul) #3

Hi Dimitris,

it looks like you are using upstart, so you could investigate the respawn stanza in your upstart script

And +1 with what Mark said you should investigate why ES crash.
ES can run with low memory, I use ES_HEAP_SIZE=1g (in ES shell script) in two VM and never had a crash

(system) #4

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