Please note: All times listed below are in Central Time Zone
Inside Puppet Labs, the SysOps team has been fighting year after year to rip up our entrenched virtualization system. Well, it is finally happening. Comprised of a very small two person team we deployed our first production grade OpenStack cloud and delivered it as a research platform to the entire company. Being what we are, a company that writes automation software and a small team with goals of using a common set of Unix/Linux tools; we decided to go it our own and build the best we could without picking up any downstream installers or commercially curated automation.
In this talk I'll go through what we gained and lost in building it ourselves, our way.
- How we overcame bugs and which architectural and implementation choices made overcoming them easier
- The skills your team will need to deploy successfully
- What planned items we had to drop from initial production release - some may have seen one of my cohort's talk's at OpenStack Vancouver titled Providing OpenStack Service High-Availability Through Anycast Routing (https://openstacksummitmay2015vancouver.sched.org/event/2qci/providing-openstack-service-high-availability-through-anycast-routing)
- How long it took us to deploy and why
Attendees will come out with a better understanding of what it still takes and why an organization implements OpenStack.