NETWAYS Blog

Day 1 – done. Awesome. Lots of interesting talks, presentations and especially those stunning demos as Dirk already remarked in his blog post yesterday. Evening event in the sky of Berlin – priceless. Well, and sleep is overrated, like usual. Personal I-need-more-time-to-play-with list: Graylog2 & Ansible in combination with Icinga 2.

OSDC Day 2 kicked off with an introduction into Docker by Tobias Schwab. While I am familiar with Vagrant used for virtualized development and demo boxes, the presentation really made it clear – give it a shot, use existing (puppet) provisioners and run containers locally on your linux. Although the project is relatively young, it’s already famous to be on the top 50 of GitHub projects. Sebastion already covered Docker already last year here on the blog. Docker further removed a blocking issue recently – the requirement of a patched kernel – in their 0.0.7 release, so definitely a „must try“ (also given that rapid development already shows 0.10.0 as latest release).
Up next a tough decision between Ceph by Martin Loschwitz, or SysDB with Sebian Harl. SysDB got my attraction in terms of its future capabilities combining it with monitoring interfaces such as the livestatus protocol for data feeds. It’s not yet decided whether to use a RDBMS/NoSQL or graph database as backend for SysDB, or how to proceed in detail with all required APIs. Though the project looks promising and possibly fills some gap on all those ‚inventory monitoring graphs metric whatever‘ tool combinations out there. Let’s wait for the documentation and a final release 😉

Automated MySQL Failover with MHA presented by Colin Charles from MariaDB/SkySQL promises insights into one those painful things to-be-done with MariaDB and high availability (and likewise, its predecessor MySQL). MHA basically monitors all connected nodes. If a failure occurs, it saves the binary logs from the crashed master. Then identifying the most-uptodate slave, applying differences to all others, promoting the new slave as new master. Last but not least, announce the new master to all slaves. Sounds pretty easy – and integrates well with existing HA solutions such as Pacemaker, as heard during the presentation.

Lunch was very delicious and sweet in the end – and so was the entry talk into the afternoon by Mike Adolphs on ‚How we do Support at Github‚. Not so much technical details, but nice images illustrating the steps how to put effort into having the best support ever. Even with Lego workers fixing all the problems 😉
Jochen Lillich with Dynamic Server Orchestration introduced alternative methods on the key player tool (like Puppet, Chef, Ansible, etc) like serf. Last but not least – the final presentation by Yves Fauser from VMWare introducing an ‚Overview of networking challenges and solutions in OpenStack‚.

Well, and it wouldn’t be me if I wouldn’t talk a lot about Icinga 2 collecting ideas & feedback during coffee breaks and on twitter #osdc. Nice location, smart people, cool atmosphere – can’t wait for next year!
Save the date – 21 – 23 April 2015!
Thanks to all speakers, sponsors and participants – see you next year 🙂
PS: We’ll upload the slides, presentation recordings and pictures in the next weeks. Keep an eye on osdc.de!