[
https://issues.apache.org/jira/browse/SOLR-11431?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16962310#comment-16962310
]
Kevin Risden commented on SOLR-11431:
-------------------------------------
I see some of these same failures locally [~psomogyi] so not sure the PR is
ready to go.
> Leader candidate cannot become leader if replica responds 500 to PeerSync
> -------------------------------------------------------------------------
>
> Key: SOLR-11431
> URL: https://issues.apache.org/jira/browse/SOLR-11431
> Project: Solr
> Issue Type: Bug
> Components: SolrCloud
> Affects Versions: 7.0
> Reporter: Mano Kovacs
> Priority: Major
> Attachments: SOLR-11431.patch
>
> Time Spent: 0.5h
> Remaining Estimate: 0h
>
> When leader candidate does PeerSync to all replicas, to download any missing
> updates, it is tolerant to failures. It uses {{cantReachIsSuccess=true}}
> switch which handles connection issue, 404 and 503 as success, since replicas
> being DOWN should not affect the process.
> However, if a replica has disk issues, the core initialization might fail and
> that results in {{500}} instead of {{503}}. I failing replica like that can
> prevent any other replicas becoming the leader.
> Proposing either:
> * Accepting {{500}} as "cant reach" so leader candidate can go on
> or
> * Changing {{SolrCoreInitializationException}} to return {{503}} instead of
> {{500}}
> * * this might be API change, however
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@lucene.apache.org
For additional commands, e-mail: issues-h...@lucene.apache.org
