Testing distributed & clustered Filesystem (Block & object based) using rpyc and other tools.

Amit05


4

Votes

Description:

Clustered or distributed Filesystems, Volume managers and other related storage softwares are not only complex to develop but are equally challenging to test. Communicating across various nodes, accessing & exploiting caches and locks across the nodes and connecting to various ports are some of the challenges every tester faces. Additionally, the resilience of these systems is another hurdle as it requires maintaining the state of test process.

Unless a robust test harness is developed, testing these products always ends in many gaps left unplugged in the code. Thankfully, we have Python. With its rich set of libaries and other related tools, ease of coding and flexibility to exploit Unix shell, tesitng scale out clusters like Hadoop, Gluster, Ceph and even commercial softwares like Veritas Cluster Filesystem has become easy like never before.

I wish to propose/demostrate/talk on how to develop test tools using rpyc and other relevant libraries to build a solid test harness.

Prerequisites:

  1. Intermediate level of Python skills.
  2. strong Unix/Linux.
  3. Understanding of Storage softwares.

Content URLs:

https://github.com/Amit05/glusterfs-automation-framework

Speaker Info:

Nearly a decade of experience in testing storage systems particularly Cluster Filesystems.

Worked at Symantec/Veritas in the File System team.

Currently works @redhat, Involved in developing new test framework - Distaf.

Explores openstack/tempest openstack/cinder , openstack/swift, ceph.

Speaker Links:

https://github.com/Amit05/python

www.amitchaurasia.com

Section: Testing
Type: Talks
Target Audience: Intermediate
Last Updated: