DEVOPS WEEKLY ISSUE #111 - 17th February 2013
This community grew partly out of the first devopsdays event in Ghent. Looking at http://devopsdays.org/ now you’ll see 14 events this year organised already, from Europe and the US all the way to Brazil, India, Tokyo and New Zealand. That’s pretty impressive. Thanks to all the people involved in organising those events, hopefully their is a good cross over between attendees and readers of this newsletter.
Devops Weekly is sponsored by Brightbox Cloud - launch cloud servers in multiple UK datacentres in seconds…
Ever wanted to use the Vagrant workflow with something other than VirtualBox? Work has been ongoing to support multiple providers for a while, and code and demos can be found with Vagrant working with VMWare Fusion. But this post, a demo of the upcoming 1.1 release, shows Vagrant working with EC2. Ideal for those times when your local machine can’t run all those virtual machines.
An article explaining why devops is needed, and predicting big inroads in traditional enterprises in 2013. The $440 million loss made by Knight Capital because of using the incorrect version of network softare might just be the best example of why good configuration management matters.
If you ever wanted a handy review of lots of the current crop of NoSQL databases then this post should come in useful. Good examples of common usecases as well as details of license, language, protocol and inner workings.
Good reminder that rapid application development needs to consider it’s approach to security, because doing it after the fact is probably too late. I particularly liked the point that with iterative development and modern technology “you risk generating a considerable amount of insecure code very quickly”.
Good talk on the link between continuous integration and deploying code to production. Lots of good tips about build time, release frequency and building feedback into the build pipeline.
Chef is pretty flexible about how you accomplish your goals, which does mean it’s possible to do things in ways that might come back to bite you later. This post covers a few of those anti-patterns, including forking community cookbooks and misuse of roles.
Kafka, the distributed publish/subscribe middleware is getting a Intra-cluster Replication capability in the upcoming release. This post describes the details, but it’s also a good read if you have an interest in the design and trade-offs involved with building distributed, fault tolerant systems.
Another good read on distributed systems, this post covers a wealth of practical hard won wisdom aimed at people new to designing and developing distributed systems. Well worth reading all the way through.
Lots more devopsdays events coming up around the world, leaving even fewer excuses to go along and meet interesting people. With Paris on 18th and 19th April, Berlin 27th and 28th of May and Sydney on July 12th and 13th. All are open for registration and are looking for sponsors and talk proposals.
GitHub have open sourced Boxen, the puppet based tool chain they use to manage OSX based development machines. Lots of examples, insights into development at GitHub and good documentation to get started with.
Fatcache says it’s memcached for SSDs. Memcached stores things in memory, which is great up until you have a huge amount of data and can’t afford the RAM. Fatcache uses a similar approach but stores things on SSDs.