[
https://issues.apache.org/jira/browse/SOLR-6453?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14146085#comment-14146085
]
Ramkumar Aiyengar commented on SOLR-6453:
-----------------------------------------
Thanks for picking this up, Noble! Just so that I understand, could you explain
the difference?
I see that the controller closes the {{Overseer}} and then the client, but
aren't we coupling the two implementations tightly by assuming that? In
hindsight, it doesn't make sense to use the {{zkClient}} when it is closed, so
I should have probably used {{if (zkClient.isClosed())}} instead of just
{{isClosed}}. Checking just {{zkController}} does make sense for the final
operation of rejoining election (which is what we are checking in the
{{finally}} block anyway), but checking just that assumes that the state reader
passed to us is from the same controller. May be we should check both, i.e. the
controller is active (if not, we can't rejoin anyway), and the client is
active..
> Stop throwing an error message from Overseer on Solr exit
> ---------------------------------------------------------
>
> Key: SOLR-6453
> URL: https://issues.apache.org/jira/browse/SOLR-6453
> Project: Solr
> Issue Type: Improvement
> Components: SolrCloud
> Reporter: Ramkumar Aiyengar
> Assignee: Noble Paul
> Priority: Minor
> Attachments: SOLR-6453.patch
>
>
> SOLR-5859 adds a leadership check every time Overseer exits loop. This
> however gets triggered even when Solr really is exiting, causing a spurious
> error. Here's a one-liner to stop that from happening.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
