search

spujadas / elk-docker

Elasticsearch, Logstash, Kibana (ELK) Docker image
Other
2.16k stars 908 forks source link

kibana service is failing in ELK 7.3. #295

Closed Venkat639 closed 4 years ago

Venkat639 commented 5 years ago

deploying latest(730) ELK in kubernetes cluster. Elasticsearch and logstash are coming up but kibana is failing with error "start-stop-daemon: warning: failed to kill 23: No such process"

tedder commented 4 years ago

Where do you see that error? I'm using 740 and kibana fails to start, but I can't find any errors. If I exec in and run service kibana start it'll start up. This is what I see n the elk-docker log:

waiting for Elasticsearch to be up (20/90)
Waiting for Elasticsearch cluster to respond (1/30)
logstash started.
 * Starting Kibana5
   ...done.

When I exec in the kibana log is empty. Of course, doing start causes it to run and the logs to fill.

tedder commented 4 years ago

Figured this out. It was failing during start too. Nothing in the logs indicating a crash (even the OS logs in the container or on the server). I finally figured it out- I needed to use the NODE_OPTIONS environment var:

NODE_OPTIONS="--max_old_space_size=4096"

While running a hacked init.d service so --verbose was used to start Kibana, I could see memory usage mostly around 540mb but it would be in the 800mb range, probably during optimize. I'm not installing any Kibana plugins so I wonder if the default size needs to be increased from 512, @spujadas.

The memory usage in Kibana is a known and open issue.

see also: #104 #105 #17 #176

spujadas commented 4 years ago

Thanks for the pointer @tedder! I've bumped up the memory limit to 4,096MB in Kibana's init file as of version/tag 7.6.0.

Closing the issue (assuming it solved the original one, if not feel free to reopen with more details).