[
https://issues.apache.org/jira/browse/SOLR-8696?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15153321#comment-15153321
]
Shalin Shekhar Mangar commented on SOLR-8696:
---------------------------------------------
Yeah but that only happens for clusters configured with legacyCloud=false.
To give you some context behind this: the default is legacyCloud=true which
means that any core admin create command can create corresponding replica in
clusterstate. This also means that if you deleted a replica which was down at
the time, it may re-create itself in the cluster state once it comes back up.
To avoid such problems, a new cluster property called 'legacyCloud' was
introduced which required all cluster state modifications to be from the
overseer only. The idea was that this "Zookeeper as truth" mode would
eventually become the default but we haven't made much progress since.
> Optimize overseer + startup
> ---------------------------
>
> Key: SOLR-8696
> URL: https://issues.apache.org/jira/browse/SOLR-8696
> Project: Solr
> Issue Type: Improvement
> Components: SolrCloud
> Affects Versions: 5.4.1
> Reporter: Scott Blum
> Labels: patch, performance, solrcloud, startup
> Attachments: SOLR-8696.patch
>
>
> ZkController.publishAndWaitForDownStates() occurs before overseer election.
> That means if there is currently no overseer, there is ironically no one to
> actually service the down state changes it's waiting on. This particularly
> affects a single-node cluster such as you might run locally for development.
> Additionally, we're doing an unnecessary ZkStateReader forced refresh on all
> Overseer operations. This isn't necessary because ZkStateReader keeps itself
> up to date.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: dev-h...@lucene.apache.org
