lucene-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Daniel Collins (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (SOLR-5071) Solrcloud change core to another shard issue
Date Wed, 24 Jul 2013 09:57:48 GMT

    [ https://issues.apache.org/jira/browse/SOLR-5071?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13718187#comment-13718187
] 

Daniel Collins commented on SOLR-5071:
--------------------------------------

I suspect the response to this will be you are doing something unsupported.  To do this properly,
you should UNLOAD the old core, and (re-)CREATE it as a new core, that would keep ZK in sync.
ZK has no knowledge of what you've done in the solr.xml file, so it can't possibly know what's
going on.

In addition, this syntax of solr.xml is deprecated as of Solr 4.4 so the whole issue would
only affect versions up to Solr 4.3.1.  From Solr 4.4 onwards, solr.xml won't have any of
this information, it will be in the core.properties file for each core.  I'm not sure what
happens if you change that file by hand, but I suspect that may to be an "unsupported" action.

                
> Solrcloud change core to another shard issue
> --------------------------------------------
>
>                 Key: SOLR-5071
>                 URL: https://issues.apache.org/jira/browse/SOLR-5071
>             Project: Solr
>          Issue Type: Bug
>          Components: SolrCloud
>            Reporter: Illu Y Ying
>         Attachments: 2013-7-24 11-55-45.png
>
>
> I have a solrcloud cluster with one collection and two shards.
> One core is a replica for shard1, I stop it and change its solr.xml like this:
> <core name="collection1" instanceDir="collection1" shard="shard2"/>
> So this core should be a shard2 replica, Then I restart it, open cloud graph page(see
attachment), you can see this core as a down replica still in shard1 and also as a active
replica in shard2.
> There is doubt about one core status in two shards.
> In this one core has two status scenario I suggest that if we could remove the down replica
information of other shard in clusterStatus.json.
> I remember when core is changing to active status, it will send overseer an active status
message, so add this logic to overseer change core to active status part.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org


Mime
View raw message