You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Jan Høydahl (JIRA)" <ji...@apache.org> on 2013/05/21 14:59:17 UTC
[jira] [Commented] (SOLR-4830) Node doesn't recover properly after
fail, when running multiple collections on same nodes with ZooKeeper
[ https://issues.apache.org/jira/browse/SOLR-4830?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13662954#comment-13662954 ]
Jan Høydahl commented on SOLR-4830:
-----------------------------------
Hi Johannes.
After discussing this issue last week I asked you to report it as a bug. But looking more closely at the config, the behavior appears to have another reason.
The solution is simply to define {{persistent="true"}} on the {{<solr>}} tag in {{solr.xml}} for each of the nodes.
What happens if persistent=false, is that the node will happily host the new collection/core initially, but will never persist knowledge about it to solr.xml. So at next startup it will still only startup the "yp" core which was pre-defined in solr.xml.
Note that the practice of persisting the list of cores in solr.xml will disappear in Solr 4.4, when a new solr.xml format with auto core-discovery is taking over.
I will close this issue, unless you still believe there is a bug somewhere
> Node doesn't recover properly after fail, when running multiple collections on same nodes with ZooKeeper
> --------------------------------------------------------------------------------------------------------
>
> Key: SOLR-4830
> URL: https://issues.apache.org/jira/browse/SOLR-4830
> Project: Solr
> Issue Type: Bug
> Components: SolrCloud
> Affects Versions: 4.3
> Reporter: Johannes Henrysson
> Attachments: pic1 - killed machine.png, pic2 - machine down.png, pic3 - end of test.png
>
>
> Created 3 collections (yp, test, hubba) with 2 shards each, on 4 nodes. So all 3 collections used same nodes.
> This worked out quite nice, until I tested to kill one node. When trying to recover, only first collection worked out. The node stayed as down on the other to.
--
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