Everyday OpenStack with Object Oriented Operations

In this extended tutorial, led by Mark Shuttleworth and James Troup, you’ll see how Canonical manages to operate tens of OpenStack clouds at reasonable scale with a small but focused team.

Operating OpenStack is widely considered to be difficult. Many early implementations have ground to a halt as people struggle with combinatorial complexity and the pace of change. Object-oriented operations enable reuse of ops code in diverse settings and configurations.

The aim of the tutorial is to walk you through several major but nevertheless absolutely standard OpenStack operational challenges. High availability deployment, scaling, ceph troubleshooting and disk replacement, and ultimately a full upgrade of a running openstack. These mechanisms are used to provide carrier grade infrastructure with regular upgrades and expansions, never allowing more than one second of network downtime.

It is an advanced talk, but people who have some familiarity with the various Openstack components will be able to follow along.