Hello, I am experimenting the K-means example (provided in the Mahout examples directory).
What I able to do: (1)I am able to run Kmeans driver and clusterdump from command line in both Hadoop pseudo-distributed mode and standalone mode. (2) I have also set up the maven plugin in eclipse and am able to run Kmeans (in mahout-core project, main class org.apache.mahout.driver.MahoutDriver, via passing "kmeans") in both Hadoop pseudo-distributed mode and standalone mode via eclipse. (3) I am able to run clusterdump(in mahout-integration project, main class: org.apache.mahout.utils.clustering.ClusterDumper) in Hadoop standalone mode via eclipse. What I am not able to do: (1) However, I *cannot* run *clusterdump* in Hadoop* pseudo-distributed mode* via *eclipse*. The exception that I got is Error: java.lang.ClassNotFoundException: org.apache.mahout.common.distance.DistanceMeasure I have added my Hadoop (pseudo-distributed mode)configuration file in the classpath for this run configuration and also tried adding all the jars search-able from the home directory of mahout (after compile/install via mvn) in the classpath of the run configuration. I am wondering could anyone give some pointers ? Thank you very much! Wei
