Bug#766233: VM won't start after an automatic upgrade to systemd inside the VM

[Tomas Pospisek]

On Wed, 22 Oct 2014, Daniel Baumann wrote:

> On 10/21/2014 11:54 PM, Tomas Pospisek wrote:
>
>> 2) release an updated lxc package for wheezy, that warns the user about
>>    the problem.
>
> i don't really see how that would be possible, without doing an abuse of
> debconf.

I'd add that to the NEWS file.

>> 3) do something in the systemd packages to work around the problem.
>
> systemd could check in postinst for being executed in a container, and
> then warn about it. this seems to be the obvious thing to do to me, but
> i guess that since you got redirected from systemd to lxc, that the
> systemd maintainers wouldn't want to do that.

I can't speak for the systemd maintainers, but when I was on the 
#debian-systemd IRC channel I had the impression that they were aching 
under heavy pressure from the migration of Debian to systemd. So I'd think 
whenever they can delegate stuff away they will.

It doesn't necessarily follow that they would refuse to include some 
reasonable solution to the wheezy-lxc with systemd-containters problem if 
offered one.

>> 4) do something during the upgrade of the sysinit packages
>
> don't think that's the right place to do it.

The sysinit packages could warn check in pre-install when upgrading to the 
version that pulls in systemd whether the system is running in an LXC 
container and then let the user know about the problem.

Is there a secure method to determine whether one is running in a wheezy 
(!) lxc container? Because if lxc 1.0.6-3 knows how to setup systemd 
containers then it'd be better to not to bother the user with a 
warning.
*t