Bug#818442: initscripts: Failing /run/shm migration and wrong message

Dmitry Bogatov KAction at debian.org
Wed Feb 20 11:36:48 GMT 2019

[2016-03-17 03:12] Pierre Ynard <linkfanel at yahoo.fr>
> During my boot sequence, when /etc/init.d/mountdevsubfs.sh is run, I get
> the following logs:

> > Filesystem mounted on /dev/shm; setting up compatibility bind mount.
> > Please remove this mount from /etc/fstab; it is no longer needed,
> > and it is preventing completion of the transition to 
> /run/shm.

> My /etc/fstab features no /dev/shm, nor /run/shm, nor any pseudo
> filesystem. So if nothing else we already have a problem in the form of
> a wrong and confusing warning message.


> Digging a little deeper, my system doesn't run udev and uses a static,
> disk-backed /dev.

It requires using own kernel, isn't it? Just tested it, and remove of
udev drags removal of initramfs-tools and kernel-image.

> It contains an empty /dev/shm/ directory. When looking
> into /lib/init/mount-functions.sh it appears it tries to permanently
> symlink /dev/shm to /run/shm, and prints the warning after it fails
> to delete my /dev/shm/ directory. The warning assumes that this is
> because it is a mount point, however it's not, and it is rather my
> understanding that the old /dev/shm/ directory removal fails simply
> because the filesystem is still read-only at that point.

Any ideas, when does it becomes read-write?
        Note, that I send and fetch email in batch, once every 24 hours.
                 If matter is urgent, try https://t.me/kaction

More information about the Debian-init-diversity mailing list