LXC Directions

[Clint Byrum]

Excerpts from Gustavo Niemeyer's message of Mon Jul 18 12:07:42 -0700 2011:
> Hey Clint,
> 
> > LXC has been something I've heard mentioned hand in hand with Ensemble
> > since the very first time I heard about Ensemble at UDS-M. I think I
> > understand the directions that have been proposed, and I think its time
> > we had the discussion so that we can enable LXC for local dev at the very
> > least, and hopefully also for many other purposes in Ensemble.
> 
> I thought we had sorted this out live during the sprint.  As we talked
> at that time, we want to use LXC in the context of service units, and
> will get the local dev story almost nailed down with that because we
> can deploy multiple units within the same machine with this already.
> Then we'll need to sort out the ability to use machine-0 for
> deployments, and finally a provider that recognizes that the local
> machine is a bit special in some regards.  We'll need to do these in a
> very piecemeal style, with very small well tested branches one at a
> time.
> 

Yes, totally, we did have that discussion, and that was the output.

I sat down to write that out as the only option in the discussion.

I'm having trouble working out how we can work inside ec2 with the network
isolation that LXC implements. After a few days trying to figure it out,
I'm proposing that, unless there's already a known solution, we move
forward with the local dev story for now.

It really boils down to this question:

How do you allocate network addresses to the containers?

The complex answer I detailed in the original post is, I think,
unacceptable.  Something simpler will have to be figured out.

I'm quite hopeful that you have something up your sleeve, or that I've
just been very dense in this, possibly blocked by my own egotistical
attachment to the work I've already done, and that its a solved problem.