by Mark Burgess
Consumer IT services continue to grow in scope and variety, and the infrastructure for bringing them to life is exploding in scale, speed, and complexity. History shows that all collaborative systems eventually fail because of the law of "diminishing marginal returns". What does this mean in IT, and how do we ensure that technology scales to meet society´s needs with a minimum of overhead? What does it mean for competitive advantage, for agilty, quality and time to market? In this talk, Mark Burgess explores more of the ideas from his latest book In Search of Certainty: The science of our information infrastructure, and explains how we can cope with necessary and sufficient complexity.
by Luke Kanies
by Adam Jacob
Devops and Continuous Delivery are two of the strongest cultural trends in Operations today, and much has been said about what exactly they entail (and many, many arguments about what they do not.) As Chief Technology Officer for Chef, I've had the pleasure of both observing the behavior of hugely successful organizations and the despair caused by businesses failing to successfully change how they operate.
In this session I'll describe how:
Devops and Continuous Delivery are reactions to the same fundamental business forces – simply solutions to the problem driven from a different point of view.
The 5 key drivers of successfully transforming your organization into a healthy, agile, forward-leaning, awesome place to work.
Why the best advice about automation projects is the worst advice in the face of a transition this huge.
The idea that culture is more important than tools is wrong, if what you care about is actually transforming your behavior.
Why you must be prescriptive about the tool-chain you are going to use, and brook no dissent once you've decided.
How to organize and run the projects that will transform your business into a model of both Devops and Continuous Delivery.
On the way we'll talk about globalization, playing guitar, configuration management, cognitive science, and the propensity of middle management to murder great ideas with the most soft and pleasant of pillows.
It'll be a good time.
How to build a continues delivery pipeline including complete test of configuration management code. I will describe how we use Jenkins to compile, create artifacts, test and verify for all our code including configuration management code. The pipeline build consist of 10 different steps that includes compilation, lint tools, Chef cookbooks test, integration test of cookbooks and application, full deployment on existing infrastructure and full the final step that is a full deployment from scratch. The pipeline is built on Jenkins, AWS, Chef, Test kitchen, lxc, Foodcritic, maven and rake. I will also show how we use feature toggling and avoid branching in application and configuration management code to be able to always have builds that are ready for deployment
by David Boucha
Open source tools for configuration management have been available for years. But a new breed of tools is expanding the realm of what is possible. These new tools are built for the speed and scale of any cloud and now provide remote execution, orchestration and deployment of heterogeneous clouds and computing platforms, real-time infrastructure data, workflow automation, and efficient, fast communications in addition to configuration management. Thomas Hatch, creator of Salt, will detail the availability and usage of various open source management tools built for the requirements of cloud at scale and help automate data center management. Configuration management is the beginnings of what is possible with open source software for infrastructure management. Attend this session to learn what more is possible.
Everything began 4 years ago, CFEngine 3 had just been released, documentation and expertise were in short supply. We had to accept the reality of a steep learning curve.
As CFEngine grew so did we. We discovered bugs, submitted pull requests, designed workarounds for various pitfalls, gained advances in productivity (thanks to the knowledge and experience gained working with CFEngine 3) and evaluated design choices available to us. This journey led us to become one of the most advanced CFEngine users in Europe.
I'll recount our journey, share insights on solution architecture with CFEngine and show examples of what we had to overcome and how we achieved that using less well-known features of CFEngine. Our examples will cover advanced use of CFEngine 3 code. Finally, I will present our retrospective: what we did right, what we did wrong and share where we have got to thus far in our journey.
by Marco Ceppi
Juju is a Service Orchestration tool designed to automate your cloud infrastructure - Configure, manage, maintain, deploy and scale efficiently with best-practice Charms on any public, private or hybrid cloud from a powerful GUI or the command-line.
We’ll look at stacks of services managed by juju and the charms behind those services. Along the way we’ll go over some of the production (flexibility, scalability, repeatability, and automation) benefits that emerge from the paradigm shift from servers to services.
This talk is designed for developers or system admins who are interested in service orchestration
by Pieter Lexis
So you've got config management, now what? Take things to higher level, try to 'outsource' simple, repetitive SysAdmin tasks you can't be bothered with ("create a DB", "Add this vhost") to your customer. But if your customer isn't a dev, DevOps scares you or you can't/won't share your code, you need to provide your customer with a way that allows you to be lazy. This will be a brainstorm-like session that discusses possible future trajectories for automating _everything_, keeping them based on config management.
by Kristian Amlie
This talk will introduce new CFEngine 3.6 features, we have these bullet points:
New language functions
by Zac Stevens
"Don't use SSH in a `for` loop!" We know we should be better than that, yet many Chef shops still rely on SSH to orchestrate their environments.
Are you using "knife ssh", but aren't happy with the security profile?
Are you using chef-solo, but wish you could use "knife ssh"?
Are you trying to invoke Chef from your other applications, and integrating with SSH feels wrong?
Solve these problems and more with MCollective, a CM-agnostic orchestration framework built on Ruby and message queues. This session shows how to deploy MCollective with Chef, integrate the two systems in various ways, use it as an ad-hoc tool, build multi-node orchestration workflows, and invoke Chef from other applications.
This session will introduce some new work. While it will include some demonstrations, this session is not a tutorial - it examines the capabilities and limitations of this approach to building orchestration, and hopes to provoke further discussion and development.
by David Boucha
Let's take 20 minutes before beginning the SaltStack training and workshop to review the agenda for the Salt room for the rest of Monday and Tuesday. We will also review an outline of the content of the just-released SaltStack Certified Engineer exam and discuss the format.
by Marco Ceppi
Charms are the real driving force behind Juju and are the instruction set on how to setup a service in a cloud using Juju.
This talk will be driven by questions and is designed for developers and system admins looking to script and manage their services in Juju.
by Michael Ducy
Sometimes you need action to happen right away, and you just need to be a bit more pushy about it. This talk till cover the features of Chef's Pushy server and how it can be used to run ad-hoc commands or orchestrate action between multiple machines. We will also cover the architecture of Pushy and give an overview of how it works internally.
by David Boucha
This workshop will begin by covering the basics of getting Salt installed on your systems. Then we'll explore the speed and power of SaltStack remote execution and cross-platform execution modules for system and infrastructure control. Once we're familiar with remote execution we'll jump into Salt states for system and application configuration management. We'll end the lab by exploring Salt Cloud and a few other advanced Salt features.
The goal of ncf is to abstract the knowledge of "how" to achieve tasks in CFEngine and expose these capabilities for non-expert users. With ncf, you can express intent in a very succinct and expressive manner, integrated with CFEngine, while keeping everything 'under-the-hood' tunable.
ncf is built from the ground-up to be easy to learn, flexible and extensible, promoting DRY-ness and knowledge sharing. It is written in pure CFEngine language, introducing a well-structured design built on multiple decoupled layers with clearly-defined interfaces and roles.
I will introduce ncf, explain how we use ncf to solve real-world problems, explain our design choices and the architectural approach taken and present a walk-through on how we build CFEngine solutions that are extensible and much easier to work with.
Razor is a flexible provisioning tool that makes it easy to control how
machines are built based on rules and policies. It maintains an inventory
of nodes and their hardware characteristics, gathered by booting each node
into a discovery image. Discovery information, together with user-defined
policies is used to make installation decisions.
Razor can install a wide variety of operating systems, from common Linux
flavors like Debian, Ubuntu, CentOS, and RHEL, to operating systems known
for their resistance to automated installation like ESXi and
Windows. Beyond installation, Razor strives to be a tool for managing a
machine's lifecycle, including power control via IPMI etc. and easy
integration with external systems.
Razor is an opinionated tool that focuses narrowly on provisioning, but
makes it easy to hand off a node after installation to a configuration
management system like Puppet to perform more complicated setup tasks and
for ongoing maintenance.This talk will give an overview of Razor's
capabilities and provide some hands-on examples about its use; it will also
give examples of how Razor and Puppet can be used to address common
provisioning problems, like building an OpenStack cloud.
by Marco Ceppi
Why and how "traditional packaging" needs to change to cater for quicker delivery. Using the strong points of OS-native packages as a tool and delivery mechanism in your CI toolchain.
How does this relate to The Foreman? Tools like fpm-cookery and/or omnibus-ruby could provide an alterantive way of delivering packages and possibly help with adopting Foreman outside of the Red Hat-focused world.
When we talk about configuration management and automation, we're often talking about automating Linux or UNIX platforms. Less frequently mentioned is how to automate Windows, an operating system that is widely used, but can often be a challenge to automate.
In this talk, I'll review some approaches to doing configuration management on Windows using out-of-the-box declarative Chef resources. I'll also discuss how Chef on Windows compares/relates to Microsoft System Center Configuration Manager (SCCM), Powershell, and some of the exciting new features in Windows 2012R2 a/k/a Desired State Configuration (DSC).
by Mandi Walls
Business continuity and disaster recovery documentation and methodologies can be onerous and consuming. Collecting information about services and requirements for a business to continue in case of catastrophe is an expensive exercise tying up employees. Configuration management tools used as a system of record for building DR plans can speed up the process as well as create a reliable and repeatable process for keeping plans up to date.
Open spaces for CFEngine-Rudder testing, bug-fixing, contributing...
With 3 CFEngine developers + 4 Rudder developers + various community members (Bas, Marco, Martin...) on hand to help, this should be a great chance to get together and test CFEngine 3.6 pre-release, squash some bugs, experiment with new or little know features, etc...
One activity will include testing the new 3.6-beta release, and posting results to an anonymous survey on https://www.surveymonkey.com/s/L....
3rd–4th February 2014