[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: Hosting & deploying OpenShift

Thanks for the answer!

I can definitely understand that a large organization might not move as harmoniously internally as outside branding might appear. And by "large organization" I think I mean, more than 5 developers :)

Re "LXC", I saw a demo of warden (currently the code is embedded in the vcap codebase) and it seems very interesting https://github.com/cloudfoundry/vcap/tree/master/warden https://github.com/cloudfoundry/vcap/tree/master/warden-client I believe it will be spun out as a standalone repo/project from a chat I had with them.


Dr Nic Williams - VP Developer Evangelism
Engine Yard
The Leading Platform as a Service
Mobile: +1 415 860 2185
Skype: nicwilliams
Twitter: @drnic

On Thursday, May 3, 2012 at 6:11 PM, Mark Lamourine wrote:

> On Thu, May 3, 2012 at 8:02 PM, Dr Nic Williams <drnicwilliams gmail com> wrote:
> I'm sorry if my reply has stopped anyone from wanting to reply. 

> Nic


We're not that easily put off Nic, :-)

I actually have a meeting scheduled tomorrow to talk to someone in the Cloudforms development and deployment group about how to implement the "Build Your Own" model using Cloudforms. (at least that's what I *think* the meeting is about).

In what follows, I'm presenting my impression and opinion based on my perspective on the process as a developer and designer.  Brett, Matt, Mike and others will certainly correct me if I get something wrong or if they disagree.

The short answer is "No, not yet"

Deltacloud, Boxgrinder, Cloudforms, Matahari and other cloud or enterprise management tools are generally distinct parallel projects.  While we do share information, (and sometimes people) during the early development phase (the first year or so, maybe more?) we don't necessarily emphasize "eat your own".  All the while we evaluate the other tech to see if/when it will be applicable. For example,  I've been watching lxc to see if we can adopt it wholesale to do some of the application encapsulation that is currently done by several distinct tools (cgroups, SELinux, file quotas).  As I mentioned I'll be discussing Cloudforms and Openshift Origin tomorrow.

Currently the production Openshift service doesn't yet rely on or use Cloudforms or Boxgrinder.  It's using mostly native AWS functions. When we started, Deltacloud was just starting and we had to get into production quickly. Since then we've been more focused on adding features and making the production service reliable and stable.  We've also replaced the 4 person core team that existed when I joined with a real ops team within the larger development group.

The other projects aren't aiming specifically at supporting us.  They have more general goals and customers in mind.  Some people might not like me expressing this opinion, but it doesn't always make sense to shoehorn marginally related projects against each other and call it a "solution". As noted above, are talking to each other and re-evaluating.  Certainly we want anything that will make our lives easier or better!

In the Openshift world, we implemented initially on RHEL on AWS/EC2, but the intent is not to lock into either the OS or the cloud service.  With the opening of the source, I expect people to port both to other OS and to other cloud platforms.  I'm guessing that we inside Red Hat will most likely concentrate on integrating the Red Hat components.  Certainly the idea of making Openshift simple to deploy using Cloudforms and BoxGrinder on libvirt/kvm or bare metal is appealing to me.

So that's the long answer: "No, Not yet" ;-)

That's my perspective and my opinion.

- Mark
Mark Lamourine <markllama gmail com>
Dad, Hubbie, Software Developer, System Administrator, Road Cyclist

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]