DEVOPS WEEKLY ISSUE #113 - 3rd March 2013
Conference season appears to have started in earnest again, which makes finding good content that little bit easier. Quite a few presentations this week, and with lots more events coming up (see the excellent list compiled by the folks from Melbourne in this issue) we can hopefully expect many more. If you see any good presentations be sure to let me know.
Sponsor
Devops Weekly is sponsored by Brightbox Cloud - launch cloud servers in multiple UK datacentres in seconds…
http://brightbox.com
News
The nice folks from Melbourne’s devops meetup, infrastructure coders, have created a list of interesting devops events from all around the world. They are looking for pull requests to keep it up to date too.
http://infrastructurecoders.com/events.html
Presentation from the recent Scale conference all about logstash from the projects creator. Mainly a good introduction, but includes a great case study from Dreamhost about improving customer support using the power of logs an dgood tooling.
http://semicomplete.com/presentations/logstash-scale11x/
Another Scale presentation, this one all about low level tools you can use to analyse the performance of linux. A rapid fire introduction to top, htop, iostat, free, ping, dstat, sar, strace, tcpdump, SystemTap, DTrace and many more. Available as both the slides and the video too.
http://joyent.com/blog/linux-performance-analysis-and-tools-brendan-gregg-s-talk-at-scale-11x
Interesting article from Yahoo about how they are working on the processing of large datasets, and how in many cases mapreduce/hadoop/batch processing don’t cut it. Some interesting dicussion of using Strom for stream processing alongside data in HDFS/HBase.
http://developer.yahoo.com/blogs/ydn/posts/2013/02/storm-and-hadoop-convergence-of-big-data-and-low-latency-processing/
Good post from Forrester about the problems with some private cloud roll-outs in large organisations. The points about putting users first make me happy.
http://blogs.forrester.com/james_staten/13-02-25-why_your_enterprise_private_cloud_is_failing
Logging is important, but can involve more moving parts than it first appears. This post covers quite a specific question, but the answer includes a wide range of handy information about file handles, pipes, log rotation and more. Worth reading.
http://nic.ferrier.me.uk/blog/2013_02/dear-nic-says-jim
A presentation from last year all about infrastructure in startups, or really about any infrastructure in young projects in any organisation.
http://www.paulhammond.org/2012/startup-infrastructure/slides/
Tools
Buildpacks are a fantastic way of extending Heroku (and soon Cloudfoundry). This one gives you your own personal ruby gems server, which might be of interest if you have private gems or you like that little bit of control.
https://github.com/hashicorp/heroku-buildpack-rubygem-server
Deliver looks an interesting (and very colourful) deployment tool based purely around Git and bash. Not needing Ruby or Python or another programming language appears to be it’s main selling point but it’s simplicity might be great for smaller projects or teams using different technologies.
https://github.com/gerhard/deliver
mcollective is incredibly powerful, but the different moving parts can make trying it out tricky. This vagrant based demo is therefore hugely welcome. By default it fires up a 6 node cluster, all configured to use mcollective and a range of plugins. The README then takes you through the different capabilities.
https://github.com/ripienaar/mcollective-vagrant