Bug#766233: LXC breakage and workarounds when upgrading VMs to jessie

[Tomas Pospisek]

Hello Marco,

On Mon, 24 Nov 2014, Marco d'Itri wrote:

> On Nov 24, Tomas Pospisek <tpo_deb at sourcepole.ch> wrote:
>
>> 1. As far as I understand this is a patch for a template used to create a
>>    new jessie LXC VM. Am I thus correct in assuming it's of no use for
>>    the "upgrade a LXC VM from wheezy to jessie"?
> Yes, the configuration of existing servers will have to be updated
> manually.

Since there is 'systemd-detect-virt', would it be possible for the 
systemd preinst to check for a LXC system and then to prompt the user 
with something like this? :

"We have detected that this is a LXC VM. When switching from sysvinit to
  systemd as is happening at this very moment, you need to make sure your
  VM is poperly configured. Otherwise it will fail to boot. Please read:

    http://wiki.debian.org/LXC/switching_to_systemd

  If this LXC VM is configured properly, then please choose PROCEED to
  continue installation. Otherwise choose ABORT and installation of systemd
  will be stopped and you can configure your VM or you can switch back to
  sysv.

  [ABORT] [PROCEED]"

Alternatively:

>> Marco d'Itri mentioned a "correctly configured VM", maybe we'll get some
>> additional input from him, on what's required in the wheezy to jessie
>> upgrade case.
> What I meant is the list of changes in the patch.

Theoreticall all that is required and that is present in the patch could 
be done automatically during systemd installation. What do you think about 
that?

Greets,
*t