cfgmgmtcamp it’s the first of the FOSDEM fringe events that started to suffer scalability issues. This year there were about 650 attendees and the biggest room available could only hold 350 people, so the organizers had to set up an overflow room.
Except for the first morning, the whole conference featured parallel tracks; related to the tools or to the methodologies. I’ve spent some time on talks related to tools I didn’t know about, just for having a rough idea of what the other communities are doing, but the track I liked the most was the main one.
Keynotes tends to be generic enough and quite inspiring; if I had to choose a talk I’ll say “Empowering developers to deploy their own data stores with Terraform and puppet” by Tomas Doran (Yelp). It was really in the mood of infrastructure at scale and gives some good advice about tools. His recipe was clear: manage the infrastructure with terraform, manage hosts with CM tools.
We know that ansible, puppet, chef, salt or CFEngine are good tool for managing machines, but do not let them handle instances creation; embrace the UNIX philosophy and use the right tool for the right job.
Half of a conference are the attendees, I met some wonderful people and has been a great pleasure to talk with them. We went out for dinner and tasted some specialties like Amadeus ribs and lots of Belgian beers.
DevOps world is evolving, we managed computers with configuration management; now we have to manage infrastructures with new tools, but also we have to plan systems capable of scaling automatically reacting to external and internal conditions.
Those who will grasp such skills will be ready for developing scalable, resilient (and hopefully successful) tools.
This is really a bold statement, especially during the first day keynotes at cfgmgmtcamp 2016 in Ghent. It sounds silly but almost every speaker suggested that the end of the CM is near, infrastructure has moved to the next level and today’s cm tools doesn’t feet very well.