To be more specific ... I can run bin/mahout seqdirectory -i <input> -o <output> and it works. If I try to run bin/mahout arff.vector -i <input> -o<output> with the same input file path, it says File Not Found. Something about that doesn't seem right ...
On Mon, May 23, 2011 at 8:10 PM, Keith Thompson <[email protected]>wrote: > I just gave hdfs://localhost/path/to/file a try but that still produced the > error (would localhost be the right thing to use there?). When I echoed > $HADOOP_CONF_PATH and $HADOOP_PATH they were both empty. Where should those > be pointing? > > > On Mon, May 23, 2011 at 8:03 PM, Paul Mahon <[email protected]> wrote: > >> If the path is for a file in HDFS, I have had problems where the hadoop >> tasks think the path is a local path, not an HDFS path. I think this >> happened to me because my HADOOP_CONF_PATH and HADOOP_PATH weren't set to a >> place that had HDFS as default. If that doesn't work, you can always try the >> fully qualified path (hdfs://server:port/path/to/file). >> >> >> On 05/23/2011 04:59 PM, Keith Thompson wrote: >> >>> I am trying to run the arff.vector algorithm but am getting a File Not >>> Found >>> Exception on my input directory. I have checked and double-checked. The >>> input path is definitely correct. Someone suggested it could be a >>> problem >>> that my Hadoop folder and my Mahout folder aren't in the same folder, but >>> I >>> was able to run the Dirchlet clustering algorithm successfully specifying >>> both the --input and --output paths so I don't think that is the problem. >>> What other problems could cause this error to pop up? >>> >>> >
