OpenStack mini conference

On 15th Feb, OpenStack India user group organised a mini conference at, Pune(India).  GNUnify was jointly organised by Symbiosis International University and Pune Linux Users Group. A mixed blend of people attended the event from students, researchers to people from the industry. The members of visiting NTT Data (Japan) OpenStack team were also present at the event. Atul Jha kicked off the event with a short introduction to OpenStack. His talk focused on OpenStack's history, its components, its supporters and also on the structure of the OpenStack Foundation. Sajid Akhtar from Netapp presented the second talk. He spoke in great detail on Storage as a Service and Cinder. His talk attracted maximum number of questions on topics like deduplication feature in swift, cinder scheduler to cinder drivers.

Kavit Munshi from Aptira spoke on using Puppet for deploying OpenStack. He described how they are using puppet to solve a lot of problems, how puppet manifests needs to be configured. Vivek Raghuwanshi from Tieto gave an introductory talk on OpenStack Networking and Kiran Murari gave a demo of OpenStack on virtualbox. One thing i learned from this talk is that giving live demo can be a risky business especially when you are dealing with SDN.

I spoke on High Availability in OpenStack. I started by explaining the complete architecture of OpenStack and describing the entire virtual machine creation process, how a call is authenticated to how messages are passed between different components and how a virtual machine gets network information when it is rendered by the hypervisor. After that i moved on towards my main topic High Availability.

HA in OpenStack is basically divided into two parts, Infrastructure HA and VM HA. I spoke on both of them in great detail like how to avoid SPOF of RabbitMQ. How we can use Active/Passive clusters for Failover HA. Other option like ZeroMQ which is based on peer to peer messaging can also be used as an alternative to RabbitMQ. After that i moved on to describe how one can make OpenStack infrastructure highly available by using Pacemaker and Corosync as a state-of-the-art high availability stack. I ended my talk by giving some insight on VM high availability in Openstack.

One of the best thing at this event which is worth mentioning was the kind of hospitality shown to speakers. Everyone of us (speakers) were surprised to see how well the event was organised by volunteers. Shout out to volunteers at GNUnify for making it an awesome event.

We have also submitted a talk on HA update for Grizzly at OpenStack Summit. Please do vote for us if you want to know more about HA in OpenStack.