Re: upgrading to jessie on container with old kernel
On Sun, May 3, 2015, at 05:47, Gunnar Wolf wrote:
> Brian May dijo [Sat, May 02, 2015 at 09:39:50AM +0000]:
> > Hello,
> >
> > If I upgrade a Jessie openvz container on my proxmox box, systemd fails to
> > start up. This is no surprise, the kernel seems to be rather old:
> >
> > root@scrooge:/# uname -a
> > Linux scrooge 2.6.32-26-pve #1 SMP Mon Oct 14 08:22:20 CEST 2013 i686
> > GNU/Linux
> >
> > Obviously, I should have set things up not to install systemd before I did
> > the upgrade. However, I want to fix this container. So, lets try:
> > (...)
> > For this particular case, when reproduced today, the solution seems to be
> > simple:
> >
> > ln -sf /bin/true /bin/systemctl
> > apt-get remove systemd-sysv
> >
> > Is this the thing I should be filling bug reports about?
>
> Hi,
>
> I know your report is WRT OpenVZ containers. I believe this to be
> useful, anyway, as the root causes are the same: I upgraded during the
> past few days a number of LXC containers (and have several more
> pending to upgrade).
>
> https://wiki.debian.org/LXC#Incompatibility_with_systemd
>
> In short, you have to install sysvinit-core in your container-based
> systems, as the host does not have the needed bits to run systemd.
That's probably not true in a general case, since I am running LXC on
jessie (LXC+systemd in host and systemd in guest) and I have been
running this with backported systemd and lxc on wheezy before jessie was
release.
Perhaps it might be a true for hosts not running systemd?
Cheers,
--
Ondřej Surý <ondrej@sury.org>
Knot DNS (https://www.knot-dns.cz/) – a high-performance DNS server
Reply to: