has anyone seen jobs hanging randomly without terminating?

[Eyal Oren]

Hi,

We're seeing some of our map tasks hanging indefinitely during execution, 
and I just wanted to check if somebody maybe had seen similar things (to 
figure out whether things are a hadoop problem or our problem).

We have a job that runs locally without problems. When pushed through 
hadoop the job hangs randomly somewhere during map execution, it doesn't 
progress but also doesn't report any exceptions (until it's killed after 
its configured timeout).  We thought it might be a concurrency problem, but 
the same happens when running the job on only one node with only one task 
running.  Also, increasing java heap to 1GB did not help. Machine is 
two-core Intel Xeon @1.86Ghz, with 2Gb of memory, hadoop 0.13.1 and Java 
1.6.0_02.

Has anyone maybe seen something similar? It's a bit hard to figure out 
what's going on since we don't get any exceptions thrown.

 -eyal