You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@mahout.apache.org by Neil Chaudhuri <nc...@potomacfusion.com> on 2012/01/07 02:03:39 UTC

InterruptedException with MeanShiftCanopyDriver

I am running MeanShiftCanopyDriver as a MR job on a Hadoop cluster, and I am getting the following exception after numerous cluster files are generated:

java.lang.InterruptedException: Mean Shift Iteration failed on input clusters/clusters-0
at org.apache.mahout.clustering.meanshift.MeanShiftCanopyDriver.runIterationMR(MeanShiftCanopyDriver.java:412)
at org.apache.mahout.clustering.meanshift.MeanShiftCanopyDriver.buildClustersMR(MeanShiftCanopyDriver.java:353)
at org.apache.mahout.clustering.meanshift.MeanShiftCanopyDriver.buildClusters(MeanShiftCanopyDriver.java:276)
at org.apache.mahout.clustering.meanshift.MeanShiftCanopyDriver.run(MeanShiftCanopyDriver.java:165)

The method call, which happens to be wrapped in a Riffle-annotated class within a Cascading flow, is as follows:

MeanShiftCanopyDriver.run(
                    configuration,
                    inputDirectoryPath,
                    outputDirectoryPath,
                    new CosineDistanceMeasure(),
                    0.05,
                    0.01,
                    0.0001,
                    25,
                    false,
                    true,
                    false);

I know this isn't much to go on, but can you provide any hints as to what the cause of the issue might be?

Thanks.