One of the best things about working at Cybera is the amount of traveling and networking that I get to do. This is very ironic, though, for two reasons. I’ve previously worked at national ISPs, law firms, and financial institutions, as well as done my own consulting. While I did travel, it was was only to get simple tasks done, and nothing else. Secondly, as a systems administrator, I’m inherently not a people person. It wasn’t until I began working for a small not-for-profit company that I found myself traveling all over Canada and the US for the sole purpose of sharing information and meeting new people – and actually enjoying it.
Three weeks ago I attended PuppetConf 2012. Puppet, quickly explained, is an open-source configuration management system that can be used to orchestrate provisioning and management of servers.
The conference was split into two main parts:
It began with the debut of Puppet’s new certification program. The first 150 or so attendees were able to take the certification exam for free — including myself. I passed the Puppet Certified Professional exam (which I guess means I can now have the acronym PCP associated with my name, without having a negative connotation). In my opinion, given the mixed reputation of IT certifications in general, I think the exam was respectable and definitely tested the relevant skills someone knowledgeable of Puppet should have.
The second part of the conference was the usual lineup of sessions and talks that you would normally see at a conference. While there were many Puppet-specific sessions, the event also had a large DevOps theme in general. This allowed people with no interest in Puppet to attend and still leave learning something new.
The majority of the sessions were recorded and are available online. Some of the best talks, in my opinion, were:
- Running Ops in EC2; Operate at scale by Jos Boumans, VP of Operations at Krux
- External Data: Hiera, ENC’s by Kelsey Hightower, Director of Operations at Puppet Labs
- Intro to PuppetDB by Deepak Giridharagopal, Director of Engineering at Puppet Labs
- We’ll Do It Live: Operations Anti-Patterns by R. Tyler Croy, Engineer at Lookout Mobile Security
- Collaboratively managing the OpenStack project with Puppet by Monty Taylor of HP
In addition to attending the sessions, I was also able to talk to people from Morphlabs, Cisco, IQT, and Puppet themselves about current Puppet trends — specifically with OpenStack.
OpenStack Grizzly Summit
Last week was spent at the OpenStack Grizzly Summit. The Folsom Summit, earlier this year, had, I believe, 700-800 attendees. This Summit had over 1,400 attendees — 200 crashed the event, as it was only able to accommodate 1,200.
OpenStack Summits are usually split into two main areas: developer design sessions and general attendee sessions. In the past, if you were not an OpenStack developer (a developer is considered someone who has contributed any amount of code to OpenStack), you could not attend the design sessions. This restriction was lifted for this Summit, which enabled general attendees to give input for future OpenStack features.
The sessions were a mix of design, 101 talks, case studies, and panels. In my opinion, there is still a severe lack of Operations-related sessions at these summits. I’m normally not a fan of presenting in front of an audience, but will look into the possibility of speaking at the next summit to help turn this around.
This is not to say the other sessions were bad, though. There were a lot of great talks and keynotes ranging from cloud interoperability, future cloud design decisions, and Quantum and Cinder (two new OpenStack projects).
Like PuppetConf, I and several other Cybera employees were able to meet with many community participants (coincidentally, everyone I met at PuppetConf was also at OpenStack). In addition, we also caught up with our friends at NeCTAR and Hastexo.
On the last day of the OpenStack Summit, Cybera skipped out in the morning to visit the San Diego Supercomputer Center.
We had lunch with Nancy Wilkins-Diehr, Chaitan Baru (a self-described “data guy”), Rick Wagner (head of HPC at SDSC), and Bob Sinkovits. We discussed differences in HPC and cloud architectures, how the two could possibly work together, and where both of their strengths are.
After lunch, Bob gave us a great tour of the SDSC data center (pictures included).
Three weeks of jet lag, airport and restaurant food, taxis, transit, hotels, and challenging myself to read Pynchon’s Gravity’s Rainbow (only on airplanes) have all been very difficult, but being able to meet and learn with members of the open source cloud and automation communities has also been a tremendous amount of fun and extremely rewarding.