Andreas Schmidt stimmt uns mit “Testing server infrastructure with serverspec” auf die OSDC 2015 ein
OSDC? Noch nie gehört…
Das ist aber schade und fast schon ein unentschuldbares Versäumnis!
Aber wir holen das nach:
Die Open Source Data Center Conference (kurz OSDC) ist unsere internationale Konferenz zum Thema Open Source Software in Rechenzentren und großen IT-Umgebungen. 2015 findet sie zum siebten Mal statt und bietet mit dem Schwerpunktthema Agile Infrastructures ganz besonders erfahrenen Administratoren und Architekten ein Forum zum Austausch und die Gelegenheit zur Aneignung des aktuellsten Know-Hows für die tägliche Praxis. Diesmal treffen wir uns dafür in Berlin!
Workshops am Vortag der Konferenz und das im Anschluss an die Veranstaltung stattfindende Puppet Camp komplettieren dabei das Rundum-sorglos-Paket für Teilnehmer, die gar nicht genug Wissen in sich aufsaugen können.
After an awesome trip to Berlin (of course with good music and no beer or was it the other way round? 😉 ), the workshops on puppet, logstash and graphite and some preparation work the conference started with our typical opening from Bernd.
For the first talk Jordan Sissel and Lennart Koopmann combined there knowledge from developing logstash and graylog2 to one talk introducing to log management. They showed all the problems you can face when dealing with logs and gave a short introduction to their tools to solve this problems. A deeper look into this tools was provided in specific talks throughout the day.
The deeper look into logstash and the complementary tools elastic search and kibana was next also provided by Jordan. One sentence every project should adapt is his philosophy that every failed user experience is a bug, a bug in software, documentation or treating people. But not also the philosophy is awesome also the feature set of the ELK stack (elastic search – logstash – kibana) so I would like to encourage everybody to try it also if he does not feel the need for log management. If you have tried it out you will get the feeling you don’t want to miss it. Only thing I hate about it is the default dark theme.
The third talk I was watching was about graphite by Devdas Bhagat from booking.com. After a short introduction to graphite he gave a deeper look into how booking.com scaled up their infrastructure and use it for monitoring and debugging. The really diverse tools that are used for collecting data and pushing them to graphite shows the power of it. Also if the common consents is that the default front-end sometimes sucks. 😉
Ansible is a configuration management solution I was pointed to quite a while ago, so I was really interested in the talk by Jan-Piet Mens, although I personally dislike the idea of using ssh for this kind of automatism. Ansible sounds really powerful and flexible but simple, especially for ad-hoc tasks. The roles provided on ansible galaxy shows the greatness of the community like puppet forge does it for puppet.
After the tasty lunch I had a difficult decision to make and decided in favor of Andreas Schmidt on serverspec. Serverspec are RSpec tests for checking servers are configured correctly. It shows that syntax does not need to be difficult and testing is simple. I would say it is simple enough to satisfy a non-technical auditer, too. Perhaps just missing a gui / report with pie charts! 😀
Lennart gave the audience a deeper look into graylog2 after that. The architectural differences between graylog2 and ELK stack are small but with big effects to consider e.g. data access. The front-end is also powerful, for dashboards not as powerful as kibana Lennart said by himself, but I think it looks nicer. I like the streams you can set acls and alerts on, too. Like always I want the good things of both (in this case graylog2 and kibana) combined! 😉
Last talk for today was Christian Patsch on capistrano and puppet. Capistrano is an orchestration tool using ssh to automatize tasks and Christian showed how to use it for a masterless puppet setup. I would not do it as a default solution but it is a nice setup if you have requirements like ad-hoc configuration. He also mentioned other solutions for those using chef or ansible as an alternative.
Now we will go to the evening event you will find some impressions from this later on and Michi will cover the talks tomorrow in another post.
Dirk ist Red Hat Spezialist und arbeitet bei NETWAYS im Bereich Consulting für Icinga, Puppet, Ansible, Foreman und andere Systems-Management-Lösungen. Früher war er bei einem Träger der gesetzlichen Rentenversicherung als Senior Administrator beschäftigt und auch für die Ausbildung der Azubis verantwortlich wie nun bei NETWAYS.