From: Cathy Sarisky (cathy_at_acornhosting.net)
Date: Mon 16 Dec 2002 - 17:35:39 GMT
Adam, which version of the kernel are you using? This is a big deal if a
user in the vserver can knock the whole box offline... especially for
those of us with less-trusted users playing with root in the vserver.
On Mon, 16 Dec 2002, Adam H. Pendleton wrote:
> At 10:44 12/16/2002, you wrote:
> >On Mon, 16 Dec 2002, Adam H. Pendleton wrote:
> > > The `vserver <name> build` operation worked great, with one problem [..]
> > >
> > > A `vserver <name> stop` operation runs [..] `/etc/init.d/network stop`,
> > > which kills all network connectivity to the box!!
> >This should fail in a vserver (or have you given the vserver more
> >capabilites like CAP_NET_ADMIN to allow it access to the kernel
> >networking interfaces that are otherwise denied)?
> No additional capabilities were given to the vserver, at least not by my
> hand. :) The /etc/init.d/network script calls /sbin/ifdown to stop
> interfaces. The interface is passed the name of the device (e.g. eth0),
> not an IP, so when executing `vserver <name> stop` the RedHat init lines go
> by, until I see
> blah blah blah [OK]
> Stopping interface eth0...
> and then it's off-line.
> > > I assume the only way to prevent this is to delete/modify /etc/init.d/
> >Best to delete them (or rather the runlevel sysvinit symlinks to them).
> >My Debian install script currently does `update-rc.d foo remove' on:
> > klogd hwclock.sh setserial urandom networking umountfs halt reboot
> >(Anything related to hardware or kernel management which is going to fail
> >anyway just sitting there and timing out).
> I will be sure to delete them, but not because they're timing out. :)
> > -Paul
> >Nottingham, GB