Thanks a lot Pat for the pointer! I tried with the latest trunk against
Hadoop 2.5.0 for the k-means cluster example, it succesfully finished the
text to tifidf vectorization MR jobs but failed at the actual clustering
complaining

 Exception in thread "main" java.lang.IllegalStateException: No input
clusters found in ...  Check your -c argument.

I had the same problem before when running against Hadoop 1.2.1, but the
problem went away after I reformatted HDFS. I didn't get lucky with HDFS
reformat this time. I'll do a bit more investigation later.

Thanks again for the pointers!

Wei


On Fri, Aug 22, 2014 at 11:20 AM, Pat Ferrel <[email protected]> wrote:

> I’m no expert here since I’m stuck on hadoop 1.2.1 but the latest master
> on github is meant to be used with hadoop 2.x as the default.
> Mahout 0.9 had some experimental support but is AFAIK not recommended for
> h 2. No clue about hadoop 2.5 specifically but if it truly is a minor
> backwards compatible release you may be ok. Mahout now uses Spark, which
> claims support through hadoop 2.4.x.
>
> On Aug 20, 2014, at 3:03 PM, Wei Zhang <[email protected]> wrote:
>
> Hello,
>
> After a system upgrade, we have a Hadoop 2.5.0 cloud instance. We are
> trying to run Mahout on top of it. We are using Mahout 0.9 (downloaded from
> http://mahout.apache.org/general/downloads.html )
>
> https://issues.apache.org/jira/browse/MAHOUT-1329 indicates Mahout
> supports
> Hadoop 2.2.0
>
> But even with mvn clean install -Dhadoop2.version=2.2.0 -DskipTests=true,
> it doesn't appear Hadoop 2.x jar is downloaded anywhere (i.e., there is
> only Hadoop 1.2.1 jar downloaded, which is what the pom.xml dictates).
>
> Further, even a simple HDFS I/O call would fail (i.e., creating an HDFS
> file). It seems Hadoop 2.2.0 is not quite compatible with Mahout 0.9
>
> My question is:
> (1) Does Mahout support Hadoop 2.5.0 or should I check out the code from
> git (as suggested at
> http://mahout.apache.org/developers/buildingmahout.html)
> in order to get a Hadoop 2.x friendly Mahout ?
>
> Thanks!
>
> Wei
>
>

Reply via email to