Hi Jeff, On 31 August 2011 22:32, Jeff Eastman <[email protected]> wrote: > No I never did. I got distracted doing something else and did not debug > further. If you are still seeing this on trunk then we should (re)open a JIRA.
I spent a bit more time on this, and the problem (at least for me) seems to me to be HADOOP_CONF_DIR. The bin/mahout script is appending it to classpath, and when it points to a legitimate Hadoop conf/ directory, something in mahout notices and switches to clustered behaviour. Could you take a look at https://issues.apache.org/jira/browse/MAHOUT-800 (and suggested patch). Does this relate to your problems? Dan > Jeff > > -----Original Message----- > From: [email protected] [mailto:[email protected]] On Behalf Of Dan > Brickley > Sent: Wednesday, August 31, 2011 12:16 PM > To: Jeff Eastman > Cc: [email protected] > Subject: Re: Problems running examples > > On 10 June 2011 18:34, Jeff Eastman <[email protected]> wrote: >> I'm still trying to figure out why reuters-0.5 does not work on either of my >> clusters. The scripts themselves have no diff and the environment variables >> are set as in trunk except for MAHOUT_HOME. The synthetic control and 20 >> newsgroups examples run on both clusters without problems (well, 20 >> newsgroups has a Version Mismatch error on CDH3, but that is another story). >> But when I run reuters on 0.5 I see "MAHOUT_LOCAL is set, running locally" >> followed by file IO exceptions in MahoutDriver that are cluster dependent. >> When I run it on trunk, I don't see this and it works just fine. > > Did you get to the bottom of this? I'm running a freshly built trunk, > compiled today with a new ~/.m2 (and I reinstalled my hadoop just in > case). Initially I was expecting the build-reuters.sh demos to run on > hadoop, but I googled errors and found my way to this thread as you've > the same errors, which suggests they'll only work locally. FWIW re > clustered mode, I have a hadoop pseudo-cluster running on my laptop > alongside mahout, and have pig and other things talking to it fine. > > I have btw had build-reuters.sh run kmeans on another -real- cluster > quite happily (and have just done that again, to make sure). Not sure > what's wrong with the Mac one, nor why > the same run of the script on my laptop tells me both "no HADOOP_HOME > set, running locally" and also "Call to localhost/127.0.0.1:9000 > failed on local exception: java.io.EOFException". From re-reading > this thread I seem to be following in your footsteps... > > Thanks for any pointers / update, > > cheers, > > Dan >
