[
https://issues.apache.org/jira/browse/SOLR-5715?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13940767#comment-13940767
]
Chase Bradford commented on SOLR-5715:
--------------------------------------
I don't see how this uses the route parameter. The list of slices that the LB
server can query is still all active slices for the collection, and not just
those that match the route.
+ DocCollection coll =
clusterState.getCollection(collectionsList.iterator().next());
+ Collection<Slice> filteredSlices =
coll.getRouter().getSearchSlices(route, reqParams , coll);
+ ClientUtils.addSlices(slices, coll.getName(), filteredSlices, true);
> CloudSolrServer should choose URLs that match _route_
> -----------------------------------------------------
>
> Key: SOLR-5715
> URL: https://issues.apache.org/jira/browse/SOLR-5715
> Project: Solr
> Issue Type: Improvement
> Components: clients - java
> Affects Versions: 4.6.1
> Reporter: Chase Bradford
> Assignee: Noble Paul
> Priority: Minor
> Fix For: 4.8, 5.0
>
> Attachments: SOLR-5715.patch
>
>
> When using CloudSolrServer to issue a request with a _route_ param, the URLs
> passed to LBHttpSolrServer should be filtered to include only hosts serving a
> slice. If there's a single shard listed, then the query can be served
> directly. Otherwise, the cluster services 3 /select requests for the query.
> As the host to replica ratio increases the probability of needing an extra
> hop goes to one, putting unnecessary strain on the cluster's network.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
