There are a large number of ways you can go down the path of getting OpenStack and it’s components installed. For my lab I chose to get familiar with a few new technologies along the way, as well as have some of these steps be something that can be extracted back into a production environment.

One more quick note before we dive in. Because of the “can extract lessons learned to production environments”, this process is being done with as many hands off, DevOps, and automated tools as possible.

Process

1. Backup – In this particular instance, I only backed up personal data. None of my prior VMware environment will be carried over. (If I need one, I can spin up AutoLab quickly).
2. Install a Razor VM – Followed the guide here to a large degree. The exceptions are documented in my prior Razor posts.
3. Set the BabyDragon’s to PXE boot – There were some oddities here due to the shape of my home lab. Basically the router I’m using won’t let me pass along the extra DHCP options needed, so, I unplugged it.
4. Set the Open Compute box to PXE Boot.
5. Assign a policy using the Razor VM to install Ubuntu 12.04 on the various boxes
6. I’ve only made it this far! Everything that follows is a rough outline.
7. Then, use Puppet to install the controller/compute bits on the Open Compute box.
8. Using Puppet, pull in the rest of the compute bits onto the remaining nodes
9. Finally, install the disks and install Ceph on the BabyDragons.

Summary

As I said in the list, the journey so far is largely documented in a few other posts (setting up Razor, assigning a policy, etc). Have no fear, other posts will be added as I get closer to laying down the actual OpenStack bits.