You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hbase.apache.org by Billy Pearson <sa...@pearsonwholesale.com> on 2008/10/08 23:12:26 UTC
TableOutputFormat
I am running jobs and I am noticing that the % done on the reduce jobs is
hitting 100% way before the job is done writing to the table
I run a map and reduce and use setOutputFormat(TableOutputFormat.class) to
write the reducer's BatchUpdates to the table I need
but on say 10M rows to update the reduce show 100% done at say a few mins
but do not complete for 30+ mins.
Any idea on why this is happening now it was not acting like this before in
0.2?
Is there a new buffer holding the reduce output or something like that?
Billy
Re: TableOutputFormat
Posted by Billy Pearson <sa...@pearsonwholesale.com>.
Appears to be a problem on the reduce task reporting in higher then 100%
done
I see task reporting as high as 377% done when viewing running task.
Billy
"Billy Pearson" <sa...@pearsonwholesale.com>
wrote in message news:gcj7nq$ubd$1@ger.gmane.org...
>I am running jobs and I am noticing that the % done on the reduce jobs is
>hitting 100% way before the job is done writing to the table
>
> I run a map and reduce and use setOutputFormat(TableOutputFormat.class) to
> write the reducer's BatchUpdates to the table I need
> but on say 10M rows to update the reduce show 100% done at say a few mins
> but do not complete for 30+ mins.
>
> Any idea on why this is happening now it was not acting like this before
> in 0.2?
> Is there a new buffer holding the reduce output or something like that?
>
> Billy
>
>
>
Re: TableOutputFormat
Posted by Ryan Smith <ry...@gmail.com>.
+1 ive seen this too.
On Wed, Oct 8, 2008 at 5:12 PM, Billy Pearson <sa...@pearsonwholesale.com>wrote:
> I am running jobs and I am noticing that the % done on the reduce jobs is
> hitting 100% way before the job is done writing to the table
>
> I run a map and reduce and use setOutputFormat(TableOutputFormat.class) to
> write the reducer's BatchUpdates to the table I need
> but on say 10M rows to update the reduce show 100% done at say a few mins
> but do not complete for 30+ mins.
>
> Any idea on why this is happening now it was not acting like this before in
> 0.2?
> Is there a new buffer holding the reduce output or something like that?
>
> Billy
>
>
>