hadoop-mapreduce-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From <Milind.Bhandar...@emc.com>
Subject Re: Is there any way for a single map job to show progress
Date Thu, 03 Nov 2011 18:22:34 GMT
By implementing getProgress().

The problem with LineRecordReader is this:

 80    if (codec != null) {
 81       in = new LineReader(codec.createInputStream(fileIn), job);
 82       end = Long.MAX_VALUE;
 83     }


And getProgress() is:

Math.min(1.0f, (pos - start) / (float)(end - start));

After the map task finishes, the progress is set to 1.0.


 *   Milind
 *

From: Steve Lewis <lordjoe2000@gmail.com<mailto:lordjoe2000@gmail.com>>
Reply-To: "mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org>"
<mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org>>
Date: Thu, 3 Nov 2011 14:02:29 -0400
To: "mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org>" <mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org>>
Subject: Re: Is there any way for a single map job to show progress

So how does a custom reader to that???

On Thu, Nov 3, 2011 at 10:28 AM, <Milind.Bhandarkar@emc.com<mailto:Milind.Bhandarkar@emc.com>>
wrote:
Individual map task progress is indicative of what percentage of input chunk has been consumed
so far by the map task. However, the responsibility of feeding this info to the framework
is the responsibility of the record reader.


 *   Milind
 *

From: Steve Lewis <lordjoe2000@gmail.com<mailto:lordjoe2000@gmail.com><mailto:lordjoe2000@gmail.com<mailto:lordjoe2000@gmail.com>>>
Reply-To: "mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org><mailto:mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org>>"
<mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org><mailto:mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org>>>
Date: Tue, 25 Oct 2011 20:55:59 -0400
To: mapreduce-user <mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org><mailto:mapreduce-user@hadoop.apache.org<mailto:mapreduce-user@hadoop.apache.org>>>
Subject: Is there any way for a single map job to show progress

My map task needs to handle a large gzipped file and site at 0% forever until it hits 100%
there is no way to split the file but it would be nice if there were some indication of progress
-
any way to do that or is progress only number of complete tasks?

--
Steven M. Lewis PhD
4221 105th Ave NE
Kirkland, WA 98033
206-384-1340<tel:206-384-1340> (cell)
Skype lordjoe_com





--
Steven M. Lewis PhD
4221 105th Ave NE
Kirkland, WA 98033
206-384-1340 (cell)
Skype lordjoe_com



Mime
View raw message