Archiv:Devops Meetup

WTF

A regular meetup of people interested in DevOps.

DevOps is kind of a new sysadmin buzzword for ways to close the gap between software development and operations; it's what happens when agile development spreads to the server room. We meet to discuss culture, processes, tools and trends that contribute to this field.

DevOps — slides from introductory lightning talk at Metaday #36

There is a mailing list for announcements and discussions, and an experimental meetup.com page.

Next meetup

Tue., 2012-07-17, 19:00, Metalab Library

Agenda

• From OpenVZ to LXC - Michael Renner and Stefan Schlesinger

Presentation/topic propositions for future

If you have something to present, discuss, show, or ask around about, but don't want to commit to a date yet – here's the place to do it. If you want to hear about something specific, add it here too - maybe there's someone here who can tell about it.

1. A comparative analysis of Linux HA Solutions by Bernhard Miklautz
2. Monitoring with Icinga (or other Stuff That Is Not Nagios)
3. …

People

A list of all users for quick reference

• Maciej Pasternacki AKA japhy <maciej@pasternacki.net>
• Michael Renner <michael.renner@amd.co.at> - More info (blog et al) on my userpage
• Hannes Schaller <admin@cyberkov.at>
• Stefan Schlesinger <sts@ono.at>
• Oskar M. Grande <musha68kBRACKETAPEgmailPUNKTcom>
• David Gabriel <b2c AT dest-unreachable net>
• Christian Hofstädtler <christian@hofstaedtler.name>
• Tobias Dostal <tobias@dostal.co.at>
• Lukas Fittl <l.fittl@efficientcloud.com>
• Dietmar Hofer <d.hofer@bikecityguide.org>
• azet <azet at azet.org>
• Michael Friedrich <michael.friedrich at univie.ac.at>

• ... you? :>

Links

• http://en.wikipedia.org/wiki/DevOps
• http://www.jedi.be/blog/2010/02/12/what-is-this-devops-thing-anyway/
• http://dev2ops.org/blog/2010/2/22/what-is-devops.html
• http://www.facebook.com/notes/facebook-engineering/making-facebook-self-healing/10150275248698920

Tools

• Automated installation & configuration
  • Opscode Chef Ruby-based automated system configuration
  • Puppet Another Ruby-based automated system configuration
  • Fai Fully Automated Installer
  • Cobbler Rapid deployment system, script/pluggable infrastructure, python-based
  • The Foreman Provisioning and external Node classification for Puppet
• Cloud
  • Eucalyptus Open-source cloud platform
  • Ubuntu Cloud
  • OpenQRM Datacenter Management Platform
• Deployment and control
  • Capistrano Remote multi-server automation tool (deployment)
  • MCollective Framework to build server orchestration or parallel job execution systems
  • Vertebra [1] Another job execution system, XMPP-based
  • ControlTier cross-platform build and deployment automation framework (job automation / infrastructure orchestration)
• Continuous Integration & Testing
  • Hudson Continuous Integration server
  • Buildbot Another Continuous Integration server
  • Selenium Web application testing system
  • Cucumber Behaviour Driven Development testing tool
• Documenting infrastructure
  • OCS-ng Inventory system + GLPI Asset Management
  • Ohai Detecting information (facts) about system
  • Facter Detecting information (facts) about system
  • puppet + facter + puppet-dashboard
  • SICE Kit A set of tools and templates designed for MediaWiki to aid people who maintain systems and infrastructure
• Interpreter "sandboxes"
  • virtualenv Python
  • rvm Ruby
  • perlbrew Perl
• Misc
  • Grml Debian-based live CD
  • Mongrel2, a language-agnostic Web server
  • Augeas, a parsing config file editor
  • IT Cockpit, a user-friendly UI for configuring Nagios

History

2012-06-19

• Aaron Zauner is going to talk about fail2ban abuse automation (see: http://pastebin.com/PvBWAhSM)
• Maciej Pasternacki has some notes about how other industries ensure reliable operations (slides, notes and links: https://lists.metalab.at/pipermail/devops/2012-June/000123.html)
• Michael Friedrich will talk about the Icinga past, present and future (notes and links: https://lists.metalab.at/pipermail/devops/2012-June/000122.html)

2012-05-29

• Continuous Integration Packaging with FPM -- japhy

• • Code snippets with comments for the build system: https://gist.github.com/2823575
  • Vagrant, a convenient command-line tool for running VirtualBox virtual machines (VMs): http://vagrantup.com/
  • FPM (Effing Package Management), the main actor: https://github.com/jordansissel/fpm/
  • Buildbot, the continuous integration server: http://trac.buildbot.net/ (alternative for open source, Travis CI: http://travis-ci.org/)
  • Articles about setting up Buildbot by Jacob Kaplan-Moss, creator of Django:
    • Part 1, http://jacobian.org/writing/buildbot/ci-is-hard/
    • Part 2, http://jacobian.org/writing/buildbot/configuration-and-architecture/
  • Runit, a sane reimplementation of Dan J. Bernstein's daemontools under BSD-like license. A tool for managing long-running services (I'd say "daemons", but they don't daemonize – they are managed by runit when running in the foreground): http://smarden.org/runit/
  • LXC - Linux Containers, a lightweight userspace virtualization for Linux (may be used for real clean builds or for testing the packages): http://lxc.sourceforge.net/
  • The Debian Policy Manual: http://www.debian.org/doc/debian-policy/index.html#contents

FPM is a command-line tool that converts between different package formats (from: local directory, ruby gem, php pear, python module, node.js npm package, deb, rpm; to: local directory, deb, rpm, solaris, tar). The "local directory" pseudo-format allows for easy creation of quick-and-dirty packages that Just Work, short-circuiting all the layers from the Debian Policy Manual all the way up to dh_make with one quick script call. FPM doesn't care about actually building your stuff (as in: unpacking, configuring, compiling, etc). It just takes directory you put it into, and turns it into a package. To prepare the compiled files and put them in the right place first is your job. So is specifying the dependencies - there are no automatic dependencies on dynamically linked libraries.

Workflow for the continuous integration using bits & pieces from the gist:

• • On developer's workstation, "vagrant up" command brings up new VirtualBox VM with clean Debian and sets up a clone of the "packages" repository with a local branch
  • "vagrant ssh" to the VM, go to the packages repo, prepare or modify new package, test it
  • "git push" from the VM to the "develop" branch on the host server
  • "git pull . develop" on the host server's master branch to merge
  • "git push" to central git repo
  • buildbot (continuous integration server) picks up the commit
  • buildbot builds changed package and adds it to the apt repo
  • new package is available to clients via apt-get within minutes from the push.

2012-02-21

• #monitoringsucks

2011-12-14

• ?

2011-11-22

1. Velocity EU recap by Michael

2011-10-24

1. Tier1 Hardware (Dell, HP, IBM) show & tell Michael
   • We're going to look at a Dell R200, HP DL380G5 as well as a IBM x3650 M3 and see how they fare in regards of hardware maintenance, configuration, out of band management and general pain points.
2. What hardware should I get anyways? General thoughts on performance evaluation, capacity planning, daily operations and other pitfalls.

2011-09-27

1. git-flow, a successful branching model – japhy
2. Bullshit Bingo: Cloud strategies (how to make cloud work for you) - japhy
3. Monitoring systems, capacity planning, trending, etc. discussion

2011-06-14

1. NexentaCore 3 ZFS Presentation & Workshop - David Gabriel

clickme: Presentation

2011-05-17

1. What's new in Perl 5.14? (Daxim)
2. Opscode-hosted Chef stuff demo: cookbook repository, platform, … (japhy)

2011-04-19

1. Kanban in operations by Tony (Datei:Kanban and Devops.pdf)
2. 5 Years Metalab discussion

2011-03-15

1. Cloud - is it good for anything? (japhy) slides

2011-02-15

1. Time management - basic ideas – japhy (based on the Time Management for System Administrators book)
2. The Foreman - cyberkov, Provisioning and external Node classification for Puppet (Presentation from FOSDEM)

2011-01-11

1. Mongrel2 introduction by Lukas Fittl
   • http://www.igvita.com/2010/09/03/zeromq-modern-fast-networking-stack/
   • http://www.igvita.com/2010/11/17/routing-with-ruby-zeromq-devices/
2. Grml fly-by by Christian Hofstädtler (slides)
3. Plat_forms - The web development platform comparison

2010-12-14

• A 10 minute PostgreSQL crash course including Q&A for operations-relevant Stuff - Michael
• A hopefully short and functional live presentation of my company's infrastructure User:b2c
  • Overview: puppet+dashboard, vmware, storage
  • live server deployment
  • automated tests
• Free talk, show&tell, …

2010-11-23 Bootstrap meeting

• WTF devops? (lightning talk, japhy)
• WTF this meetup? (discussion)
• Achieving Fully Automated Provisioning (whitepaper review and discussion, japhy, sauce: http://blog.controltier.com/2009/04/new-whitepaper-achieving-fully.html and http://dev2ops.org/blog/2009/11/2/6-months-in-fully-automated-provisioning-revisited.html)