Issue 36 - 28th August

Anyone thinking of coming along to the next devopsdays event in Goteborg in October might want to know that the early bird ticket price is due to end on 1st of September. And anyone still thinking of submitting a talk idea needs to do so before the 15th of September. See everyone there.


Great presentation from Froscon by Kris Buytaert entitled Devops, Devops, Devops. It’s another nice introduction to the topic and a call to get involved in the discussions.

Interesting blog post on what is termed here the configuration management divide, partly focusing on getting developers and sysadmins collaborating on config management, and why the deployment story sometimes gets in the way. I have some thoughts and code on this I need to write up.

An article by Luke Kanies of Puppet Labs which, as well as being a nice introduction to devops, reveals some of the results of a recent survey on the subject.

When and what to virtualise is a question that needs dealing with in most environments, and this article covers a few basic issues, along with some recommendations and opinions. Worth reading if it’s something you’re new to.

Not everyone chooses to use a puppetmaster when using puppet, and this article covers a simple approach to kicking of puppet runs using Fabric.

Lots of the videos from Devops Down Under are now available. They include talks on Kanban for IT Operations, feature flipping, building high-performing teams and more.

Amazon continue adding popular services to those available in EC2, this week rolling out Elasticache. It’s an in-memory cache which talks the memcached protocol, so if you’re already using memcached instances on EC2 it should be simple to try this out.

It’s fair to say the explosion of ops tools hasn’t really hit the shores of those working in a Windows environment. This article takes a look at this issue and outlines why some Windows only tools right now would be a good step forward and make some of the problems easier to manage.

Have you ever needed to spread execution of commands across a few servers? Well this article looks at that problem with 10,000 servers and shows a few approaches, with code and graphed results, of how one team approached the problem.

A nice in-depth article on managing your personal workstation using Chef, written by a self confessed newbie to all things configuration management. It’s actually a pretty good introduction to some of the Chef moving parts too.

Interesting piece on the benefits of continuous deployment. It’s a little focused on engineer driven startups but still covers some of the reasons why investing in the tools to deploy this way lead to more important cultural changes too.


Following on from the puppet lint tool from last week, this week I came across some rspec matches for testing puppet modules. After quite a bit of talk about testing configuration management code we now have quite a range of tools to choose from.

I like the idea of applications exposing information about themselves over http, and Insight looks a nice dashboard for displaying this sort of information.

Lots of folks have been getting interested in node.js, and quite a few of the tools I’ve featured in this newsletter over the weeks have used node. Anyone put off by javascript though might find node.x of interest. It’s basically a node port to the JVM, currently with examples in Jruby and Groovy with more planned.

If you’re using VMWare fusion but looking wishing for a nice command line interface then fission might be exactly what you need.