LOADays 2013 and Kumina

Yesterday evening we returned from Loadays 2013 in Antwerp after two days of talks, meeting interesting people and Belgian beers. Being there allowed us to see how our peers do their administrating, what new technologies are on the rise and some general presentations on related subjects (like this awesome vim overview). In this blogpost I’ll go through some things we learned and want to implement in the (near) future.

Jan Piet Mens gave a very convincing presentation (with a lot of puns) on Ansible. Ansible relies on the existing SSH infrastructure to deploy system configuration, it is written in python, extensible and has very few dependencies. Right now it’s Monday morning and we’re already busy writing code to deploy Apache ServerAliases via a web interface and Ansible.

Right now we use KVM with custom scripts that use libvirt to deploy virtual machines. However, as the amount of virtualization hosts grows, it becomes harder to get an overview of what is going on. We want to start testing some cloud-like solutions. Thankfully on LOADays there were some presentations on these cloud frameworks. In the coming months we want to test (and perhaps migrate) to either OpenNebula, OpenStack or CloudStack. And in the long term, splitting up the storage infrastructure from the virtualization infrastructure using perhaps a ceph cluster.

One of the most awesome things we saw was a presention on ExtreMon by Frank Marien, an extensible, ‘extreme’ monitoring system that provides realtime, very-high precision data and allows the creation of custom filters to show (graphically) the most important data, create live overviews for it and zoom in for more detailed information.

So while we’re trying out all these new technologies, expect blogposts on the progress we make, the problems we encounter and howtos when we finally implement this technologies.

Tags: ansible, ceph, cloud, extremon, load2013, loadays, monitoring