You are viewing a plain text version of this content. The canonical link for it is here.

Posted to oak-issues@jackrabbit.apache.org by "Marcel Reutegger (JIRA)" <ji...@apache.org> on 2013/09/18 17:13:52 UTC

[jira] [Commented] (OAK-1023) Optimize many child nodes access

    [ https://issues.apache.org/jira/browse/OAK-1023?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13770859#comment-13770859 ] 

Marcel Reutegger commented on OAK-1023:
---------------------------------------

One specific problem is with the getNodes() method when an offset is used. This will immediately cause the MongoMK to read all child nodes. Most of the time the intent of the client it not to read all nodes, but just the next N nodes. MongoMK should in this case just read N + offset nodes instead of all nodes. Better yet, the implementation should be able to actually leverage the offset properly. However, this is not possible with the MicroKernel API and the underlying content model we have with MongoMK. MongoMK would require the name of the last child node instead of the offset to efficiently get the next batch of child nodes.
                
> Optimize many child nodes access
> --------------------------------
>
>                 Key: OAK-1023
>                 URL: https://issues.apache.org/jira/browse/OAK-1023
>             Project: Jackrabbit Oak
>          Issue Type: Improvement
>          Components: core, mongomk
>    Affects Versions: 0.9
>            Reporter: Marcel Reutegger
>
> Access to many child nodes is still somewhat slow with MongoMK. One area where this surfaces quite quickly is the UUID index, where all the UUIDs are stored as sibling nodes under the :index parent.

--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira