[
https://issues.apache.org/jira/browse/SOLR-8349?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15156315#comment-15156315
]
Gus Heck commented on SOLR-8349:
--------------------------------
I do see how what you suggest could be made to work if a decoder was supplied
and the decoded version was what was cached.
So are you proposing changing JarRepository, or adding a similar class called
BlobRepository? JarRepository does seem very Jar focused. (returning
JarContent, JarContentRef, etc..) It would for example no longer be possible to
return a ByteBuffer type from getByteBuffer, since returning the bytes from
which to decode a HashMap won't help. Seems like fairly major surgery is
required to make the JarRepository class fully generic.
I do like the versioned resource and config API bit though, that's nice.
I need to better understand the lazy="true" bit you mentioned, because I very
much don't ever want a user query to be the thing that causes the resource to
load (a process that could take minutes). The resource should be available and
ready to go when the system starts accepting queries.
> Allow sharing of large in memory data structures across cores
> -------------------------------------------------------------
>
> Key: SOLR-8349
> URL: https://issues.apache.org/jira/browse/SOLR-8349
> Project: Solr
> Issue Type: Improvement
> Components: Server
> Affects Versions: 5.3
> Reporter: Gus Heck
> Assignee: Noble Paul
> Attachments: SOLR-8349.patch, SOLR-8349.patch
>
>
> In some cases search components or analysis classes may utilize a large
> dictionary or other in-memory structure. When multiple cores are loaded with
> identical configurations utilizing this large in memory structure, each core
> holds it's own copy in memory. This has been noted in the past and a specific
> case reported in SOLR-3443. This patch provides a generalized capability, and
> if accepted, this capability will then be used to fix SOLR-3443.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
