At work we try, and sometimes even succeed, in using Test Driven Deployment so as one of my background projects I’ve been wrapping certain tools in to cucumber friendly forms. Over the last couple of days I’ve been grabbing ten minutes here and there to incorporate Puppet 2.6 in to the pile. Feature: Puppetwrappers Puppet Provider Examples Scenario: Confirming package installation When a machine has been puppeted Then the bash package should be installed Scenario: Confirm doodoodoo package is absent When a machine has been puppeted Then the doodoodoo package should not be installed Scenario: Confirm cron service is running When a machine has been puppeted Then the cron service should be running Scenario: Confirm tomcat6 service is not running When a machine has been puppeted Then the tomcat6 service should not be running Scenario: Confirm dwilson is in libvirtd group When a machine has been puppeted Then dwilson should be a member of libvirtd Scenario: Confirm dwilson has a uid of 1000 When a machine has been puppeted Then dwilson should have a uid of 1000 Scenario: Confirm dwilson has a given shell When a machine has been puppeted Then dwilson should have the /bin/bash shell I really like using the puppet providers for this because of the abstraction benefits they provide. Read on →

Tonights (the March 2011) London OpenSolaris User Group (LOSUG) was a little different to usual and while the topics have always been quite diverse we’ve never had as seditious a talk as one covering the Solaris fork, OpenIndiana, Illumos and the OpenSolaris community. Alasdair Lumsden did an excellent job of explaining the new projects, why they exist and what they’re aiming for. As someone who took a few steps back when Oracle purchased Solaris it was an interesting catch up. Read on →

It all started with one of those annoying little items on the todo list - find all the unpuppeted ssh authorized_keys files on a machine and alert on them. On first impressions it was going to be quite manual (always a bad sign), involve digging in to legacy installs and would be something we’d need to re-verify occasionally. It couldn’t be that bad though could it? After all how many places can an unmanaged-by- puppet sshkey live? Read on →

One of puppets more under-appreciated features is its ability to abstract and smooth the edges of certain operating system tasks and behaviours. Even something as trivial as installing a package can actually become a portability nightmare once you consider the number of different systems in the wild - rpm, yum, dpkg, pkgsrc etc. - and the varied commands needed to use them. You end up either hard coding commands, and sacrificing portability, or writing your own detection, lookup and invocation logic. Read on →

It might be a sign that I spend too much time online but the quicker a system gives me feedback the more useful I find it. While I love knowing my Nagios safety net has me covered when making changes sometimes waiting for that cgi to refresh can take too long, especially if I’m taking a iterative / test driven approach to the changes I’m making. For those use cases I wrote nrpe-runner. Read on →

While adopting a configuration management tool like Chef and Puppet will have a large, nearly immediate, effect on your work flow even after using the tools for a while you’ll still get a little smile at all the little niceties you continuously discover. One recent small win we had recently was bringing some apache configs files under Puppet command. When we started we had the following block of config: RewriteCond %{REMOTE_ADDR} ! RewriteCond %{REMOTE_ADDR} ! RewriteCond %{REMOTE_ADDR} ! It’s not hard to read and roughly understand what it does, but you have no real context; magic numbers keep things terse but are rarely the most helpful when in the land of a strange system. Read on →

Between Xmas and New Year I had some spare time to invest on a side project I’ve been looking forward to working on for quite a while. I’m pleased to announce the opening of the Puppet CookBook. I’ve introduced Puppet to quite a few companies, sysadmins and development teams over the years and a lot of the same issues, concepts and needs repeatedly crop up. By explaining how puppet works in terms of tasks and desired outcomes rather than in raw feature descriptions I hope to show some of its power and flexibility in easy to use examples in a different way to most of the existing documentation. Read on →

Hadoop is one of those technologies that seems to have forever changed the way parts of the industry work but has had no effect on my actual job. In an attempt to keep myself current for the after techtalk conversations I decided to buy Hadoop: The Definitive Guide, Second Edition by Tom White - and I’m very happy with the choice. While there are massive amounts of information online about Hadoop and the ecosystem emerging around it I still found HadoopTDG to be a useful book and worth the money (especially on the iPad as it’s a bit big for comfortable tube reading). Read on →

Over the years I’ve had the opportunity to attend a lot of different events focused on quite a few different programming languages, but none of them match the sheer enthusiasm and love of the language that you get from London PM. While there is always a contingent of LPMers at Perl conferences held further abroad the London Perl Workshop is my yearly chance to see lots of old friends, what they’ve been up to and discuss what’s coming next in our field. Read on →

I’ve been watching the Marionette Collective for a while, and even gave it a small trial in a couple of testing environments, but this weekend was the first time I’ve experimented with it at a slightly larger scale (just over a hundred small VM nodes - you have to love EC2) and I’m still impressed. I can see how it’s going to make parts of my work flow easier, and in an attempt to learn a little more about how the plugin system works under the hood I decided to write a small agent, FileMD5er. Read on →