David Smiley created SOLR-10597:
-----------------------------------
Summary: Overseer election endless loop on NodeExists
/overseer_elect/leader
Key: SOLR-10597
URL: https://issues.apache.org/jira/browse/SOLR-10597
Project: Solr
Issue Type: Bug
Security Level: Public (Default Security Level. Issues are Public)
Components: SolrCloud
Affects Versions: 6.5.1
Environment: Docker/Kontena
Reporter: David Smiley
Somehow, my 5-node Solr cluster got into a state in which _each node_ reported
it was going to be the leader and then failed to create /overseer_elect/leader
since it was already there. This generated tons of logs without end. It's
hard to say what other detrimental effect it had but Solr seemed available. To
resolve the problem, I shut the cluster down (except Zookeeper), then brought
it back up.
In case it matters: in Docker/Kontena, host name to IP (DNS) can change
frequently.
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
