[
https://issues.apache.org/jira/browse/SOLR-1477?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12762844#action_12762844
]
Jason Rutherglen commented on SOLR-1477:
----------------------------------------
I made some headway with getting this working. I pass a tier
parameter that is incremented when the request is recognized as
being distributed in handleRequestBody. This way if we're on the
2nd tier, the system can skip retrieving docs on the 1st pass.
The issue is, QueryComponent.createRetrieveDocs on the 2nd tier
needs ShardDocs, however the information for SD isn't available
on the 2nd tier so it needs to be passed in from the 1st tier.
The concern is that there's a lot of parameters to pass to
recreate a sharddoc when it's only to obtain document data.
I wish the distributed wasn't so bound together where objects
are modified on the request and response ends. It's not
intuitive. If we expect people to write or customize distributed
enabled components maybe we need to think about a cleaner API.
> Search on local cores
> ---------------------
>
> Key: SOLR-1477
> URL: https://issues.apache.org/jira/browse/SOLR-1477
> Project: Solr
> Issue Type: Improvement
> Affects Versions: 1.4
> Reporter: Jason Rutherglen
> Priority: Trivial
> Fix For: 1.5
>
> Attachments: SOLR-1477.patch, SOLR-1477.patch, SOLR-1477.patch,
> SOLR-1477.patch, SOLR-1477.patch
>
>
> Search on cores in the container, using distributed search.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
