Accidental DevOps day at DevNation and Red Hat Summit
There is only one day left for Red Hat Summit and DevNation, and my brain is about at capacity for what it will be able to remember! I had a most excellent day yesterday (Wednesday) and I am so proud of my teammates. They rocked their presentations!
The amusing and frustrating part about the day was that I had 2-3 presentations picked out for each time slot, and it was very hard to make a decision on where to go. Is “eeny, meeny, miny, mo” an appropriate way to select talks? I also did not intend to get three talks into a day that all had something to do with DevOps!
In addition to the talks I went to, I spent a good bit of time at the Partner Pavilion today giving the vendors some love. I also attended the Women in IT Leadership luncheon with a fantastic talk by Dr. Caroline Simard. If you don’t know who she is, I recommend looking her up.
And on to the track notes!
Meeting the challenge with PaaS: OpenShift in the enterprise with Tom Benninger, Anderson Silva and Andrew Butcher
Best DevOps message of the week goes to Tom Benninger. 🙂
“Working with people, specifically IT people, specifically operations people can be tough. Self-service models can help smooth the way… OpenShift removes many many steps between you and your application’s deployment.” –Tom Benninger
Red Hat IT’s path to a production deployment of OpenShift:
- IT tested the alpha version of OpenShift Enterprise
- Established an architecture & operating model
- 90-day company-wide proof of concept
- Revised the architecture & operating model
- Production release with SLAs and externally accessible applications
And the success?
- In 30 days after production release, over 130 new apps were deployed. Today, Red Hat IT’s installation of OpenShift runs over 500 apps. <– Woot!
- They were able to deploy a Heartbleed detector application in 16 hours (end-to-end process) after the announcement. The detector application is up on access.redhat.com/labs if you are interested.
How Red Hat IT works towards continuous improvement with OpenShift:
- Our internal Customer One program – as new products come out, Red Hat IT tries to be the first customer to provide usability feedback to our product teams.
- Using the OpenShift Community – critical to helping Red Hat IT in understanding how to use the tool, and also to get enhancements we need to provide better support
Presentation notes are located here.
If you take nothing away from reading this, take this one thought: The OpenStack team is concerned about the release-to-release model and wants to move towards learning about things they need to fix before the changes are landed. Or, in other words… the current distribution model is not supporting the concept of rapid prototyping and early feedback loops. DevOps. 🙂
Interesting facts for the OpenStack team:
- Over 2,000 developers on the project
- 40-50 patches uploaded an hour
- Use 10-20k new clouds a day based on the patches submitted
- Velocity is not a problem
- They have two types of users – users who consume the major releases, and users who want continuous delivery (commit-to-commit updates)
- They have a third party testing model – interested parties can set up a testing system, hook into OpenStack’s downstream and can report back if a new update will break their eco-system
Key things to ponder from the presentation:
- Motivation to be in a continuous delivery model is contribution. You want to contribute.
- Software development isn’t always the end of the process; there are also considerations for marketing, user documentation, sales, etc.
- If the line stops, everyone should stop and focus on fixing the issue.
- Develop a healthy respect for testing and code review. “If it’s not tested, it’s broken.”
DevOps message: Think about your enterprise software that can’t go out frequently because your customers would hate you if you updated it so often. How do you get early feedback loops with your customers so you know you are working towards the right thing?
Tim works on my team in Red Hat IT, so I feel like part of my goal of attending was to discover content that I could use at a future point to troll him. Instead, I witnessed a rocked-out presentation and some of the best audience participation I’ve seen at the convention.
Tim and Andrew talked about Red Hat IT’s year long EAP6 migration and some tips and tricks on how to do it smoothly. I think I would do a disservice to Andrew & Tim’s presentation if I tried to regurgitate what they talked about. I recommend taking a look at their presentation. I very much appreciated the mixture of “this is what JBoss can do with you” with the “practical application from Red Hat IT.” It was a very effective way to communicate the message (and I’m not just saying that because I’m bias. Promise!)
DevOps message: Handle your JBoss deployments with automated scripts. No manual deployments!
Amusing note: If you are curious about what the room looked like, Tim took a picture of the crowd at the top of the presentation for his mom.
Join the Red Hat Developer Program (it’s free) and get access to related cheat sheets, books, and product downloads.