jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jukka Zitting (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (JCR-2989) Support for embedded index aggregates
Date Fri, 17 Jun 2011 14:44:47 GMT

    [ https://issues.apache.org/jira/browse/JCR-2989?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13051101#comment-13051101

Jukka Zitting commented on JCR-2989:

The recursive behaviour I'm worried about is due to the way aggregates are built by the indexer.
For example say that you have a subtree of 1m nt:folders organized in three levels (100 child
nodes for each non-leaf folder). Now, given the above example aggregate definition, whenever
any one of these folders gets updated, the search index will also trigger an update of the
index entry of the top-level folder. The SearchIndex.createDocument() call would then have
to recursively traverse the entire subtree of 1m folders just to build the correct new index
entry. That's a performance nightmare just waiting to happen.

Recursive aggregates can be a pretty useful tool and I'd certainly like to see them fully
supported, but I'd like to avoid a situation where it's too easy to shoot yourself in the
foot with careless configuration.

> Support for embedded index aggregates
> -------------------------------------
>                 Key: JCR-2989
>                 URL: https://issues.apache.org/jira/browse/JCR-2989
>             Project: Jackrabbit Content Repository
>          Issue Type: Improvement
>            Reporter: Alex Parvulescu
>            Assignee: Alex Parvulescu
>            Priority: Minor
>         Attachments: JCR-2989.patch, JCR-2989_v2.patch
> Index aggregates could contain other index aggregates. JR should be able to handle a
complete hierarchy of aggregates. 
> I'm working on a patch.

This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira


View raw message