[
https://issues.apache.org/jira/browse/CHUKWA-30?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12709499#action_12709499
]
Jerome Boulon commented on CHUKWA-30:
-------------------------------------
Theoretically, yes, in practice it's not currently possible.
I have access o the FSOutputStream so I can call flush and it should have just
flush to disk except that the HDFS implementation is not doing anything on the
flush instead there's a flushbuffer method that write to disk but that one is
private or protected ...
I gonna see with the HDFS team if there's a way to flush without too much
overhead because the flushbuffer also compute the crc. If that's possible then
we can extends the class or play with introspection to get access to this
method ... but yes, I want to be able to flush after every post so I'll work on
that direction.
> Remove HDFS flush & connection holding (Collector)
> --------------------------------------------------
>
> Key: CHUKWA-30
> URL: https://issues.apache.org/jira/browse/CHUKWA-30
> Project: Hadoop Chukwa
> Issue Type: Improvement
> Components: data collection
> Reporter: Jerome Boulon
> Assignee: Jerome Boulon
> Attachments: CHUKWA-30.patch
>
>
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
