Re: /run in experimental
On Friday 13 May 2011, Roger Leigh wrote:
> On Fri, May 13, 2011 at 09:07:04PM +0200, Stefan Lippers-Hollmann wrote:
> > On Friday 13 May 2011, Roger Leigh wrote:
> > > On Wed, Apr 27, 2011 at 05:40:13PM +0100, Roger Leigh wrote:
> > > > On Sun, Apr 17, 2011 at 10:23:10AM +0100, Roger Leigh wrote:
> > Switching to /run/ triggers a critical bug in ifupdown's init script
> > (#623076, #623523, proposed change in #625204). This completely breaks
> > any kind of network access (including lo), if /etc/network/run is a
> > symlink to /dev/shm/network (default).
> Is this definitely the case with initscripts 2.88dsf-13.6 /after/ a
> reboot? It's working fine for me:
Sorry about the noise, I apparently didn't re-test 2.88dsf-13.6
throuroughly enough, after this bug triggered with 2.88dsf-13.[3-5].
It does indeed work after a reboot, which moves the relative symlinks
to absolute ones.
> # ls -l /etc/network/run
> lrwxrwxrwx 1 root root 16 May 13 20:15 /etc/network/run -> /dev/shm/network
> # ls -l /etc/network/run/
> total 4
> -rw-r--r-- 1 root root 16 May 13 20:17 ifstate
> # cat /etc/network/run/ifstate
> # ls -l /dev/shm
> lrwxrwxrwx 1 root root 8 May 13 20:17 /dev/shm -> /run/shm
> ls -l /run/shm
> total 0
> drwxr-xr-x 2 root root 60 May 13 20:17 network
> Note that 2.88dsf-13.6 switched the /dev/shm symbolic link to being an
> absolute link (/run/shm, not ../run/shm). But you'll need to reboot
> (or rerun mountall.sh after deleting the link) to fix it up--it may
> well be the relative link that's causing problems for ifupdown.
Yes, this seems to avoid the bug in ifupdown and keeps the network
> > Wouldn't it be better to upload a NMU for ifupdown incorporating the
> > proposed modifications first and adding a versioned Breaks to
> > initscripts 2.88dsf-13.6 accordingly?
> Certainly if this is still broken. If you could confirm the above
> for me, that would be great. If it is, we'll fix ifupdown first.
This doesn't appear to be necessary anymore, as the absolute symlinks
in sysvinit >= 2.88dsf-13.6 don't trigger ifupdown's problematic
behaviour anymore, thanks a lot.