[Lxc-users] Can't start containers

[Troy Telford]

On 2012-04-02 12:58:43 +0000, Serge Hallyn said:
> (You've probably mentioned this before, but) can you tell us exactly what
> host distro+release you're on?  Any customizations?

For myself, I'm using Debian (sid), no customizations; just the stock 
distro kernel (and packages)

> I'd like to get to the bottom of this, but do not have time *right* now.
> If you need to work around this, you could do worse than that simply
> replace the pivot_root(new_root, old_root) call with chroot(new_root)
> and get rid of the subsequent umount(old_root).

That's a useful tip; thanks.

In my case, it was working (very well) with the Debian kernel packages 
from Jan 2012 - linux-image-3.2.0-1-amd64_3.2.1-1_amd64.deb seemed to 
work just fine for me.

While I suspect something kernel-related (since it was a reboot that 
changed my system from working fine to its current state), I'm just not 
sure what the cause could be. Is this a kernel issue, or userspace?

The thing that blows my mind is that liblxc *vanishes entirely* from 
the system's dynamic linker after I try to run lxc the first time.

As a follow up of other things I've tried:

* Creating a "new" lxc container:  Won't start, same errors.
* Starting a previously "working" container at boot: Won't start; I 
don't recall the error offhand, but I do know the container had no 
issues before, and was shutdown properly.

Another note:
	- Debian defaults to using /var/lib/lxc. Not wanting to go to the 
trouble to change this, I did a bind mount in my fstab:
	/srv/lxc /var/lib/lxc none bind 0 0

I don't see a reason the bind mount should cause a problem, but I 
figure it's worth mentioning.
-- 
Troy Telford