Open Stack

Hi All,

I am trying to figure out high availability and was wondering if
anyone here has some good reading material on the subject?

I currently have 4 physical machines (each with 8 dual-core 3GHz CPU
and 128 GB of RAM) but we plan to scale up to a larger deployment in
the near future of about 16 physical machines.

Only two of the machines have connections to the outside over 1GbE so
I am planning on making those the cloud controllers.   Seeing each
machine is very capable I would also like to VMs on the two nodes
designated as the cloud controllers.  Is this a bad idea?  I could
also run them at a reduced load.

Here is what I am thinking:
https://docs.google.com/open?id=0ByOuh2TRmUHgcnI4ZU5UZTR3OU0

Also does openstack have the ability to automatically re-start VMs on
a new node that were running on a failed node?

This is pretty new territory for me so any information you can send my
way would be greatly appreciated.

Thanks,
Sam