Devops Weekly A weekly slice of devops news
Issue 22 – 20th May 2011
More articles worth reading, events worth attending and tools worth investigating. If only I had time to try out all of the things everyone is publishing or discussing.
The next Devopsdays Europe has a date, Friday 14th and Saturday 15th of October 2011, to go along with the location of Goteborg in Sweden. They are still looking for sponsors so if that’s you be sure to get in touch with the organisers.
Pro Puppet is out. James Turnbull has written a follow up to the first Puppet book, this time with much more content and with an index (sorry James). It covers getting started all the way through to using MCollective. Apologies for the affiliate link, it helps pay the bills.
This is the first PDF link I’ve included I think, but it’s worth it. The Operating Systems Division of The National Security Agency has published a whopping 200 page guide to the Secure Configuration of Red Hat Enterprise Linux 5. If this is a topic you need to know about this looks perfect.
I’m going to wager lots of readers spend a great deal of time in the console, and why not when so many powerful tools exist all just a keystroke away? This post is collecting useful command line tools, a few of which I’d not come across before. Particularly good starting point if you’re new to the command line.
Everyone is probably vaguely (or intimately) familiar with the Unix Process model. Adam Wiggins of Heroku has been writing about applying this approach to web applications, in particular as an approach to scaling up a system. The tool mentioned called Foreman looks pretty interesting.
A nice long form article by Jez Humble aimed at CIOs and technical managers. Introducing the ideas behind continuous delivery and other devops tenets, and focusing on the business rationale to make changes.
A nice article arguing that, yes, adopting devops practices will probably lead to more work due to unfulfilled demand, but that this is a good thing. And with the right tools, people and processes in place should both drive down costs and drive up profits. Good examples to backup the theme too.
It’s easier to build apps upon apps with advances in APIs and simpler to fire up new servers in cloud environments either inside or outside your company. This proliferation and speed makes traditional business continuity planning harder argues this article. Getting ahead of this potential problem is something worth thinking about for any growing organisation.
Having system metrics is important, but without application level monitoring your users can have problems even while you think your servers are fine. This article walks through a great example with graphs and diagrams showing exactly how this comes about.
FPM saw a few more releases this last week, amongst other things adding the ability to very easily convert Python packages from PyPi or local directories to DEBs or RPMs.
Splinter looks a interesting Python based web functional testing tool. It looks a little like Capybara in Ruby, and provides three drivers for chrome, firefox and the python zopetestbrowser. The blog post demonstrates using it with Lettuce which is a Python BDD framework similar to Cucumber.