Well, that is my problem here: i checked the logs on resourcemanager and
nodemanagers and saw nothing suspicious, but these logs are too verbose.
Could it be the data or avro records which have issues ?
--
Alexandre Fouche
Lead operations engineer, cloud architect
http://www.cleverscale.com | @
This usually happens when some task fail, their progress is then not
counted, hence the 'restart'. Check your task logs for failures.
Jan
On Wed, Nov 7, 2012 at 12:30 PM, Alexandre Fouche
wrote:
> I have a Yarn MR (with two ec2 instances to mapreduce) job on a dataset of
> approximately a thous
