Chris, thanks for your quick response. The only artifact ID I am using is:

HADOOP = artifact('org.apache.hadoop:hadoop-client:jar:2.0.0-mr1-cdh4.0.1')

and the repository line:

repositories.remote <<
'https://repository.cloudera.com/artifactory/cloudera-repos/'

(both taken from my buildfile, see my previous message's attachment
for the full thing).

In CDH4, dependencies are handled by using the hadoop-client artifact
rather than hadoop-core. This is because CDH4 comes with both MRv1 and
YARN, and thus why I need to set the classpath to
"/usr/lib/hadoop/client-0.20/*". This is explained fully in the CDH4
docs here:

https://ccp.cloudera.com/display/CDH4DOC/Managing+Hadoop+API+Dependencies+in+CDH4

I haven't had much luck finding example buildfiles for Hadoop on the
web...it doesn't help that Google thinks that "buildr" is a
misspelling.

Chris

Reply via email to