I was trying to access Kibana and I've got redirected to page with following message in red:
plugin:elasticsearch@6.2.4 {"code":"ECONNRESET"}
Kibana's log:
kibana-re-custom | {"type":"log","@timestamp":"2018-06-19T14:57:11Z","tags":["error","elasticsearch","admin"],"pid":1,"message":"Request error, retrying\nGET http://esc1:9200/_nodes?filter_path=nodes.*.version%2Cnodes.*.http.publish_address%2Cnodes.*.ip => socket hang up"}
yet when I looked at elasticsearch log, there is no errors and it seems to be working just fine. I restarted both of them elasticsearch and kibana and everything is back to normal.
As I stated earlier after restart both Elasticsearch and Kibana everything is back to normal, however I would like figure out what exactly went wrong and what can I do to prevent this in future or at least do better troubleshooting to help you guys figure out an actual cause...
It sounds like you might be experiencing the same problem that another user reported. His solution was to put a proxy in between Kibana and Elasticsearch, but an easier solution for you might be to set the elasticsearch.healthcheck.delay property to a higher value (see linked issue).
I'm more than willing to experiment with elasticsearch.healthcheck.delay value, however I can't find what's default value and how high should I increase it too?
You're absolutely right, setting up a proxy is outside the scope of that setup document. I wouldn't advise going down that route right now, until we've exhausted other options.
The elasticsearch.healthcheck.delay defaults to 2500, which is 2.5 minutes (it expects milliseconds). I would start by doubling that to 5000 (5 minutes), and see if that alleviates your symptoms. If not, try bumping that up to 8 or 10 minutes. If you're having issues after that, then there could be something else going on
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.