Paul, the trace is from the task log at logs/hadoop/userlogs/<job>/.....
Sean: I'm investigating. I don't rule out it can be Brisk related though. All I'm running is: mahout org.apache.mahout.clustering.syntheticcontrol.canopy.Job On Fri, Jun 24, 2011 at 11:54 AM, Sean Owen <[email protected]> wrote: > OK well what input/output is it reading from? it's either missing, or isn't > the right format, or is corrupted, or maybe it's the _SUCCESS thing again. > It's got to be something like that. I don't know -- can't see what you're > doing from here. But would be good to know if you spy a problem. > > 2011/6/24 Patricio Echagüe <[email protected]> > > > I see. Unfortunately all the trace I see is the one I pasted. > > > > It looks like it is writing to a local file when it fails. >
