In 2013, we saw the introduction of many new DevOps tools and existing tools such as Vagrant saw some huge improvements and features added. With all this change, its hard to keep up with not only what has changed, but how to use these changes and what benefits they have.
This talk begins by briefly introducing a handful of my own tools: Vagrant, Packer, and Serf. I talk about what their purpose is, some of their latest features, and who is using these tools. I then continue to show how these tools can be used along with other software such as Docker to create some truly amazing development and production stacks. For production stacks, I discuss the benefits of using some of these tools over existing options, and for development stacks, I show how Vagrant improves process and workflow to help streamline the entire process from development through to production.
Your mother might be cooler than mine, but security is hard work. Heavily automated environments like “the cloud” bring both new strategies and new challenges. We’ll work through some real-life problems and solutions; topics maybe your mom forgot to mention. Off-topic topics will include: controlling and tracking network access; 2-factor authentication and you; security-minded logging and metrics, and the importance of using protection for data at rest. There will be code.
better tooling for building consistent servers. With the introduction of Packer this last year, it’s become easier to automate the process of building consistent images. Thinking of images as compiled (immutable), configurable, runnable, containers opens up some interesting possibilities.
In this talk I’ll be introducing the Immutable Pipeline, a process our team uses to ensure consistency between staging and production. I will also be covering ways we test our images before they move to production.
I was a college-student BOFH, an ISP sysadmin, and an academic IT manager. If I wanted a curmudgeonly sysadmin, I could just look in a mirror. Most developers were unlikely to be granted root on any systems whatsoever, and servers were carefully hand-whittled works of art. You know, The Way Things Were.
Moving to a role doing operations at a developer-centric startup was both a culture shock and a great learning experience. I’ll discuss what worked for me (and what was challenging!) in terms of my transition to a DevOps practice. Spoiler alert: change is scary, but awesome developers on the team make it much easier for an old-school sysadmin to be assimilated into the DevOps of Borg.kromh
Docker is an open source LXC-based container service that was released in March 2013. It makes it easy to create lightweight, portable, and self-sufficient containers. Containers which you can use to test applications, build and run services or even to build your own platform-as-a-service. Learn why Docker matters, how to get started with it and see some cool examples of Docker in action. This talk will explain:
Getting started with Docker
Demo some cool Docker use cases
It’ll also discuss the right places to use Docker and try to answer some of the questions around using Docker with other tools like Puppet and Chef. By the end of the talk you’ll see both how Docker is useful and how to make use of it.
SaltStack does a lot of stuff, mostly focusing on infrastructure automation, configuration management and cloud orchestration. Quite simply, our software helps automate DevOps tasks for speed and scale on the Internet assembly line. But there’s much more, so Tom will provide a quick overview of the SaltStack platform and share a few things you probably don’t know about how SaltStack is used in DevOps organizations today. Tom’s talk will highlight the need for speed and scale in DevOps, put an end to an age-old holy war in configuration management, talk about why we need more lightweight cloud control, outline the benefits of automation glue for all the infrastructure things, and show how test automation can be easier and faster combining Salt, Jenkins, Docker and Git.
We use Chef to manage a large and rapidly growing infrastructure at Shopify. Our cookbooks are open to development not only by our Operations Team, but by all 200+ developers at the company, and we manage to maintain an environment of continuous delivery and continuous improvement for our infrastructure. I’ll talk about some of our workflows, internal tooling, and security considerations when operating at our scale.
DevOps is a fascinating, quickly growing field. With Chef, Puppet, Ansible, Salt, Packer, Docker, Vagrant, AWS, and others, there are many ways to get an infrastructure up and running. However, it is quickly evident to those entering this field that choosing one tool often excludes all others or requires significant investment in custom development to fit the chosen tool(s) into the intended infrastructure. In this talk I’ll be introducing Pantry, a tool-agnostic provisioning service we’ve been building and running at Collective Idea. Pantry’s job is to facilitate communication between developers and their infrastructure, ease the burden on setting up new infrastructure, and to allow seamless use of multiple DevOps tools together. I will be talking about how Pantry came to be, the ideas and technology it uses, the problems Pantry solves for us at Collective Idea and what the future will bring.
Typical Puppet users implement this tool in its traditional role of client and master. At Adaptive Computing we build environments for developers all day long using OpenStack, as such the traditional model of client to server for Puppet just didn’t fit. For the longest time I thought I was doing it wrong, however Puppet can be implemented without the need of a puppetmaster. The presentation will focus on utilizing puppet in a masterless state allowing the rapid configuration for your environment with no need for the cruft of managing a traditional puppet master. There are pros and cons that will be explored as well as detailed examples of how Adaptive took environment installs from hours of developers time down to about 10 minutes.
Has “DevOps” jumped the shark?
Some say yes; others say 2014 will be the year DevOps dons its Fonz-esque leather jacket. Whichever you believe, the marketing feeding frenzy has begun and the dilution of the “DevOps” concept to include everything (and simultaneously mean nothing) is palpable.
But DevOps has a dirty secret: many of its core tenets are ancient computer history: configuration management, version control, automated builds, and metrics-driven decision making: all old news. So what caused these long-known best practices to congeal and pop together, becoming part of the industry consciousness? And why have some been so successful at implementing it in their organizations while others continue to struggle?
This talk deconstructs the meta-elements of DevOps that made it resonate so strongly with so many and allowed those familiar DevOps poster children—Netflix, Etsy, and others—to deploy the methodology with such success in their businesses. We’ll go beyond DevOps’ classical CAMS (culture, automation, metrics, and sharing) definition to discover what exactly what made DevOps relevant, and what about it is so timeless and foundational that it will make whatever-follows-DevOps relevant, too.