nutch-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Andrzej Bialecki (JIRA)" <j...@apache.org>
Subject [jira] Commented: (NUTCH-799) SOLRIndexer to commit once all reducers have finished
Date Fri, 05 Mar 2010 09:56:27 GMT

    [ https://issues.apache.org/jira/browse/NUTCH-799?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12841790#action_12841790
] 

Andrzej Bialecki  commented on NUTCH-799:
-----------------------------------------

I think it's ok to do it this way - the commit per reducer may be actually harmful if commit
succeeds but the task is killed for any reason and re-ran.

Note: the patch has some formatting errors.

> SOLRIndexer to commit once all reducers have finished
> -----------------------------------------------------
>
>                 Key: NUTCH-799
>                 URL: https://issues.apache.org/jira/browse/NUTCH-799
>             Project: Nutch
>          Issue Type: Improvement
>          Components: indexer
>            Reporter: Julien Nioche
>             Fix For: 1.1
>
>         Attachments: NUTCH-799.patch
>
>
> What about doing only one SOLR commit after the MR job has finished in SOLRIndexer instead
of doing that at the end of every Reducer? 
> I ran into timeout exceptions in some of my reducers and I suspect that this was due
to the fact that other reducers had already finished and called commit. 

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message