Hi all, I have installed Hadoop 0.20.2 on a cluster of 6 computers. Everything runs fine when I run an application on Hadoop with one system in the cluster. But as soon as I include more than one system and then run a job, I get the following error after completion of the map task:
Shuffle Error: Exceeded MAX_FAILED_UNIQUE_FETCHES; bailing-out. The value of "ulimit -n" is set to 35000. I am not using any DNS names, so no DNS resolution problems. What can be the problem then? Same happens with Hadoop 0.21.0 version. Thanks, Prashant.
