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

Posted to hdfs-user@hadoop.apache.org by stack <st...@duboce.net> on 2009/10/07 23:02:54 UTC

Thanks for the super excellent, fancy-pants sync/flush/append

Thanks hdfs'ers for getting in the sync/flush (Where do I send the check?
Smile).

I'm currently trying to rig hbase so that this feature is on in our TRUNK so
all can help test the new feature.  It doesn't look like I can get to hflush
from the SequenceFile.  I should rig my own version local down in hbase?

Thanks all,
St.Ack

Re: Thanks for the super excellent, fancy-pants sync/flush/append

Posted by Hairong Kuang <ku...@gmail.com>.

Temporarily you can continue to use the method "sync" in FSDataOutputStream.
In HDFS, "sync" will eventually call "hflush". Because of the project split,
we can not make "hflush" visible in FileSystem before the append-branch is
merged with the hdfs trunk.

Hairong

On Wed, Oct 7, 2009 at 2:02 PM, stack <st...@duboce.net> wrote:

> Thanks hdfs'ers for getting in the sync/flush (Where do I send the check?
> Smile).
>
> I'm currently trying to rig hbase so that this feature is on in our TRUNK
> so all can help test the new feature.  It doesn't look like I can get to
> hflush from the SequenceFile.  I should rig my own version local down in
> hbase?
>
> Thanks all,
> St.Ack
>