FWIW after searching for the same libraryI had to build spark it to get it to work with HDFS on a cloudera install
I downloaded the CDH version from the spark site and still had to build it to get it to work, this is the command I used
SPARK_HADOOP_VERSION=2.0.0-cdh4.6.0 sbt/sbt assembly
SPARK_HADOOP_VERSION=2.0.0-cdh4.6.0 sbt/sbt publish-local
Iíve run both Spark-Shell remotely and applications linked against the built library
The Maven central repo contains an artifact for spark 0.9.0 built with unmodified Hadoop, and the Cloudera repo contains an artifact for spark 0.9.0 built with CDH 5 beta. Is there a repo that contains spark-core built against a non-beta version of CDH
(such as 4.4.0)?