[
https://issues.apache.org/jira/browse/SOLR-6001?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14602132#comment-14602132
]
Arcadius Ahouansou commented on SOLR-6001:
------------------------------------------
Hello [~elyograg], I need a feature that looks very similar to this one.
Let me explain:
Currently in our setting, we have let's say 3 SolrCLoud nodes on servers A, B
and C.
The 3 nodes are identical by design and have exactly 1 collection ( with only 1
shard, 1 replica) each.
On each of these servers, we also have running a read-only API client that uses
the CloudSolrClient in solrJ to query Solr.
Since all solr nodes are identical, it would be interesting to see how we could
tell the API to by preference query its local SolrCLoud node first i.e we need
to allow for SolrJ a setting like
{code}
-Dsolr.preferredReplica=solrOnServerA
{code}
or something along this line.
And other replicas will be queried if and only if the local node having a copy
of the collection becomes unavailable.
Note that we have a separate indexer which will not use this config for
indexing data.
This seems to be something that could be implemented in SOlrJ only, so it may
not be quite what this ticket is about.
Thanks
Arcadius.
> Local/Leader shard/node preference for SolrCloud
> ------------------------------------------------
>
> Key: SOLR-6001
> URL: https://issues.apache.org/jira/browse/SOLR-6001
> Project: Solr
> Issue Type: Improvement
> Reporter: Shawn Heisey
> Labels: solrcloud
>
> It would be useful if SolrCloud had special preference options on queries for
> preferring local nodes or shard leaders.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
