Deploying PySpark on Red Hat Storage GlusterFS

by Steve Watt, Chief Architect, Big Data, Red Hat

Red Hat and Continuum Analytics are pleased to announce a new solution that allows customers to deploy PySpark on top of Red Hat Storage GlusterFS. If you're attending Strata, you are encouraged to swing by the Red Hat Booth to grab a solution brief that describes how the solution is put together and how you can set it up. However, for those of you that are not at Strata, here's the overview -- and be sure to check out the technology brief, here.

Continuum Analytics are the makers of Anaconda, a leading Python distribution. At Strata, Continuum Analytics are announcing a new product, Anaconda Cluster, which is a highly-scalable cluster resource management tool. Red Hat Storage GlusterFS is a cost effective, easily scalable, POSIX compliant, distributed filesystem that runs on industry standard servers. Given that accessing data in HDFS from Python can be cumbersome, Red Hat and Continuum Analytics have built a solution that enables Anaconda Cluster to deploy PySpark on GlusterFS. This collocated solution keeps life simple for Python developers by providing a Python interface to Apache Spark that is able to read and write data on a distributed filesystem that looks and works like the local filesystems that they are used to. Furthermore, given that both Python and GlusterFS are written in C, this allows easy access to data from Python applications whether they are running on-premise or in the cloud.

If you'd like to try it out, please check out the demo posted in the video below and its accompanying tutorial: https://github.com/wattsteve/pyspark-tutorial