You may be better off considering a different approach all together.

We ran into similar issues involving FISMA constraints and data
classification in the past.  One
'simple' solution is complete segregation ( not to be confused with
isolation ).  What we ended up coming up with was to basically deploy one
openstack cluster as an operations control cluster then use it to deploy
other openstack clusters as a service.

If you tie that into live migration you can even move physical machines
around between nodes by shutting them off and re-provisioning them as a new
node entirely.

The primary issue in this approach is network automation.

One solution is to deploy gateway systems as a service, but that's a bit of
a hack.

If you are interested in discussing this further I'd be glad to go into
depth in a blog post or something for yah.


On Mon, Jun 11, 2012 at 8:41 AM, Christian Parpart <> wrote:

> Hi all,
> while I am still "somewhat" new to OpenStack, I was able to successfully
> deploy a 6-node OpenStack setup
> with 4 compute nodes, one controller node (yet to be HA'd) and one network
> node (yet to be HA'd).
> However, now, that I am that far, I am in need to create a custom filter,
> due to the companies requirements
> on what VMs to put on what hardware.
> Unfortunately, I am not yet that experienced in Python (know quite a few
> others, so I shouldn't have
> it that hard in getting into) and I - of course - don't know the Nova API
> as well as you do.
> So I am looking for a complete basic Hello World Filter Scheduler example,
> e.g. in form of a github repo,
> that I can fork off, and improve it, and learn Nova by doing.
> I could not really find anything that helpful yet, and I think it might be
> a wonderful entry-point for
> quite a few of us.
> Is there anyone willing to help us here out a bit?
> Many many thanks,
> Christian Parpart.
> _______________________________________________
> Mailing list:
> Post to     :
> Unsubscribe :
> More help   :
Mailing list:
Post to     :
Unsubscribe :
More help   :

Reply via email to