1

I have cloudera CDH5 running inside a virtual box.

when I try to run :

mahout spark-itemsimilarity ....

I get the error:

Unknown program 'spark-itemsimilarity' chosen.

Do i have to install any additional package to run the spark-similarity?

Any help would be appreciated !

CrazyBrazilian
  • 143
  • 3
  • What version of Mahout are you using? I think that Spark job execution supported was provided since Mahout 0.10. – Shagun Sodhani May 18 '16 at 09:47
  • The output of 'rpm -qa | grep mahout' is 'mahout-0.9+cdh5.7.0+29-1.cdh5.7.0.p0.79.el6.noarch'. So I guess Cloudera CDH5.7 comes with Mahout 0.9. Is that right? – CrazyBrazilian May 18 '16 at 23:48

1 Answers1

1

Spark support for Mahout came from Mahout 0.10 release while you are using 0.9 release. So this should explain why you get the unknown program error. I would suggest using a higher version of Mahout.

Shagun Sodhani
  • 722
  • 4
  • 26
  • Cloudera only ships with 0.9+ however Mahout doesn't require complicated configs. Simply download Mahout and make sure `SPARK_HOME` is set properly in the env variables and it should work. (better yet- please call your CDH rep, and tell them you want Mahout 0.13.0) – rawkintrevo Jul 16 '17 at 04:51