You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@lucene.apache.org by "Alan Woodward (JIRA)" <ji...@apache.org> on 2016/10/19 15:42:58 UTC
[jira] [Updated] (SOLR-9664) CloudSolrClient's cluster state cache
can break query routing for new shards
[ https://issues.apache.org/jira/browse/SOLR-9664?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Alan Woodward updated SOLR-9664:
--------------------------------
Attachment: SOLR-9664.patch
Here's a patch with a test case illustrating the problem. It creates a collection, adds some docs via a CloudSolrClient (populating the local cache), and then adds a shard. Queries routed to the new shard will fail because the local cluster state hasn't been updated.
[~noble.paul] any suggestions here?
> CloudSolrClient's cluster state cache can break query routing for new shards
> ----------------------------------------------------------------------------
>
> Key: SOLR-9664
> URL: https://issues.apache.org/jira/browse/SOLR-9664
> Project: Solr
> Issue Type: Bug
> Security Level: Public(Default Security Level. Issues are Public)
> Affects Versions: 6.3
> Reporter: Alan Woodward
> Attachments: SOLR-9664.patch
>
>
> CloudSolrClient can do query routing if a query has an explicit route parameter set on it. This doesn't play well with the cluster state cache, however, because the routing is done entirely client-side, with no opportunity for a node to report that the cache is out-of-date.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@lucene.apache.org
For additional commands, e-mail: dev-help@lucene.apache.org