When we installed Elastic and Kibana on ou Kubernetes environment we have the recurrent issue that on a restart the index-pattern creation fails. It's an old issue that doesn't get solved in the latest release of Elastic.
The sequence to reproduce the issue :
restart Elastic nodes , restart Kibana
data is ingested into indexes
goto index-management to create index patterns
create index-pattern , example : apm-*
goto discover in Kibana => no index pattern found !
restart Kibana process on Kubernetes separately , without restarting Elastic
Hi Matthias,
We use RedHat linux. There is indeed a reverse proxy involved. Not sure if this makes a difference. If the caching of the proxy would be playing a role then the restart of Kibana shouldn't make a difference ? The phenomenon only presents itself on the first index pattern created after a full restart, starting completely empty on indexes, dashboards,..
More specifically we deploy on REL-8 with docker images from Elastic which contain CentOS Linux version 7.
Lieven
Hi
Thanks for those details, I'm collecting evidence that there's an issue, you're not the first one reporting this. The previous reports were all using a proxy. Mind if I ask, are there any logs entries in the kibana log or in the log of your reverse proxy that could point to an issue?
Thx!
Matthias
Hi,
since it's only presenting itself at a new restart,I'll have to check the next time. Being part of a big company, we don't control the proxy ourselves. Any idea what kind of symptom we should be looking for ? My impression is that the root cause is kibana and elastic going out of sync on the list of index patterns. Checking if we have the previous logs of Kibana before restart...
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.