I’ve spent the last week on holiday in the lovely city of Budapest and at the Craft conference on software craftsmanship. Once the videos start coming out I’ll hopefully link to them as there was lots of great talks, and it was interesting to be at an event with very few operations people for a change.

Sponsor

Devops Weekly is sponsored by Brightbox Cloud - serious UK-based cloud infrastructure from only 1.5p per hour (£10.95/month)

Start your £20 free trial now: http://brightbox.com/devopsweekly

News

A pretty complete walk through of lots of the moving parts in CoreOS, including using etcd and confd for configuration and using bringing up a cluster on EC2 using CloudFormation.
http://marceldegraaf.net/2014/04/24/experimenting-with-coreos-confd-etcd-fleet-and-cloudformation.html

One of my favourite talks at the recent Craft conference was this one on Software Craftsmanship. Lots worth taking to heart, from the importance of API design to learning from research in our area.
https://speakerdeck.com/postwait/the-paradox-of-software-craftsmanship

My own talk from Craft, on continuous integration for infrastructure. A set of patterns and tips for testing virtual machine images, automating load testing and writing acceptance tests for provisioning systems.
https://speakerdeck.com/garethr/continuous-integration-for-infrastructure

Lots of systems are now distributed, with lots of parts joined together by a network. This post does a good job of explaining why this is different to an application running on a single machine. Lots of hard won tips aimed at people just getting started, but probably useful for everyone.
http://www.somethingsimilar.com/2013/01/14/notes-on-distributed-systems-for-young-bloods/

An interesting idea to formalise a schema for metrics. The site contains a little of the how, along with a formal specification and links to a few existing tools.
http://metrics20.org/

Every wanted to learn SELinux but thought it looked complex? Well this next link is for you. It’s a colouring book which explains how SELinux works. Yes, really.
https://github.com/mairin/selinux-coloring-book

A nice simple example of making monitoring more usable by linking alerts to entries in a runbook. Usability tips and tricks like this are really useful.
http://riltsken.github.io/devops/infrastructure/monitoring/2014/04/19/making-runbooks-more-useful-by-exposing-them-through-monitoring.html

A call to stop using Jenkins and similar for continuous deployment pipelines. Although I agree with the sentiment I’ve yet to see anything massively better, at least in the open source space.
http://www.cloudsidekick.com/blog/stretch-armstrong.html

If you’re convincing Windows administrators and developers of the benefits of infrastructure tooling (or anything) it’s often useful to have something from Microsoft to point to. This new site links to lots of Windows specific tutorials and software releases under the devops banner.
http://msopentech.com/blog/project-categories/devops/

Here’s a nice example of using Ansible to orchestrate a zero-downtime deployment. Specifically we’re taking instances out of a load balancer, deploying our application, re-adding to the load balancer and then moving on to the next one.
http://developer.rackspace.com/blog/rolling-deployments-with-ansible-and-cloud-load-balancers.html

Events

If you’re interested in devops adoption in very large organisations you will likely find the Devops Enterprise Summit pretty interesting. October 21st to 23rd in the San Francisco Bay area.
http://www.devopsenterprisesummit.com/

Tools

There are a few tools cropping up recently that aim to make configuration of load balancers more dynamic. Frontrunner works with Marathon and Mesos only, but it’s an interesting way of configuring haproxy.
https://github.com/Wizcorp/frontrunner

As mentioned a few weeks ago the Go continuous delivery tool is now freely available, and as of this last week the source code is now open source. Will be interested to see if a community builds around this.
https://github.com/gocd/gocd