Sensible Ansible: AnsibleFest 2016 in London
Author: David Negreiea
Elements’ dynamic duo David and David traveled to London last Thursday February 18th to attend AnsibleFest 2016. DevOps engineer David Negreira recaps on the event in this blog post.
Currently Elements is using Ansible inside the company in order to make the deployment of back-end applications faster and easier. It made all the sense that David Negreira and David de Sousa, who are pushing forward the utilization of Ansible at Elements, to attend the event. It makes even more sense specially if you consider that after the recent news of RedHat acquiring Ansible and the launch of Ansible 2.0, we were super curious what the future will bring for the development of Ansible and how people are making use of the nice new features.
Ansible Core team
Todd Barr, Tim Cramer, Greg DeKoenigsberg and Bill Nottingham started the first two talks of the conference. They have shown that in the last two years the number of contributors have been growing and the amount of work that five members in the core team of Ansible are now dealing with more contributions of more than 2,000 developers. This is one of the points where Redhat is coming to help and organize how the internal structure of development will be defined in the near future. We can also expect more predictable release cycles from the team, meaning that we can also plan internally when a new release is out and what new features we can use.
Network device support
Peter Sprygada did a presentation on the newly announced support for network device vendors like Juniper and Cisco. This will allow enterprises to make us of Ansible to configure the whole network stack, giving more control over the infrastructure that an application or service is part of.
Ansible and ILM
After a coffee break it was time for Aaron Carey and Jim Vanns from Industrial Light and Magic to demonstrate how they deployed Mesos on their infrastructure with Ansible. This was a nice tech talk full of tips and tricks where they showed some of their playbooks and explained how they were doing deployments to multiple cloud providers. The best trick we learned was that we can use the Ansible uri module in order to send HTTP REST calls to any service, really an eyeopener which has shown us some extra capabilities that we can make use of.
One-on-one sessions
After this talk, and as we had been given the opportunity, we scheduled a 1-on-1 session with one of the experts at Ansible.
We had the opportunity to speak with James Laska and pose some questions that we had on using Ansible Galaxy. It turns out that there is no standard or correct way of using Ansible Galaxy and developing it internally as we want. We were advised to explain our use case and send our feedback as this is an internal process that is ongoing in the Ansible Galaxy project. It has no definition yet as they are still gathering the feedback from their users. We, as Elements, will happily provide our feedback and help moving the open source community further with this discussion.
After the questions left out in the open, time to check the sponsors, get a demo from DataDogs team and refill energy with a nice lunch.
Writing modules
The afternoon started with a presentation from James Cammarata on writing modules for Ansible. He has shown the best practices in order to develop modules and even made a demonstration (code included) on how to control the Philips Hue lights bridge using Ansible with his own module for Ansible. It was a very funny and curious use case which I would never expect to see from Ansible.
Ansible and Atlassian
After James’ talk it was time for Steve Smith, from Atlassian to demonstrate how we can make use of Ansible to provide continuous development on their products. Pretty nifty as it allows a bigger way of controlling and knowing what is happening, integrated with JIRA and Stash and all their other tools, all automatically done from their web interface. We learned that we can control and verify which version currently sits on the different environments and how we can make use of ITIL principles to make sure that there are the necessary approvals and information in place before deploying new code.
Accelerate deployments at Société Génerale
The next talk was done by Fabrice Bernhard, CTO at Theodo, who explained some tips and tricks on how to change the mentality of a big company in order to accelerate deployments at French multinational bank Société Génerale. He also talked how Ansible had a crucial role in the whole process, as It allowed for developers and operations people to sit around and read and understand the simple code that Ansible is made of, accelerating time from coding to deployment and basically their whole development process.
DevOps best practices
After the afternoon coffee break it was time to have a DevOps best practices interactive discussion with some of the Ansible people, not much to take from this discussion to be honest.
Cisco and Ansible
Next up was Fabrizio Maccioni, technical marketing engineer at Cisco, who talked about the work that Cisco is doing along with Ansible in order to configure and template the configurations of network devices. He also talked about how this will help companies to add the network stack to their deployment scripts in the future, as the network is still on an island on its own when it comes to deploying a new application. He also went a bit deeper on how they are working with Ansible and working on an Open Format in order to be able for their devices to be configured in a standardized way. It is very good to see that Cisco is helping these great open source projects and contributing back to the community.
Deploying with Docker
Last talk of the day was given by Vik Bhatti, senior platform engineer at Beamly, and he explained how they are deploying Docker and building AWS images to deploy immutable infrastructure and how he advocates that everyone should be using it. We got some really new concepts and ideas out of this talk, we will look further in the future to see how can we make use of immutable infrastructure for our projects.
Conclusion
Overall we got a really good impression of the current and future state of Ansible. We also understood that we are keeping up at the same pace as other companies, facing the same issues and solving the same issues the same way. Good to see confirmation that Elements is following the right path keeping up with the latest technology and pushing its barriers.
Originally published at www.elements.nl on February 19, 2016.
