Hi Am 22.04.2018 um 15:09 schrieb Antonio Terceiro: > # findmnt | grep machine-id > ├─/root/patched2/etc/machine-id /dev/mapper/lemur--vg-root[/root/patched2/run/machine-id] ext4 ro,relatime,errors=remount-ro,data=ordered > > This explains the crash in mkosi and points the problem to something > that happens during the debootstrap run. > > I compared the output of debootstrap from unstable with the patched > debootstrap, and they are idential, i.e. packages are installed in the > same order, but for some reason, when running with the faster > debootstrap, the above mount is left over. > > Looking around, I suspect that this could be left behind by > systemd-machine-id-setup, however, I couldn't understand yet why this > would happen. > > systemd team: could you provide any insight? for reference, I am > attaching the current diff between debootstrap master branch, and a > local branch where I have Thomas Lange's patched applied. I just stumbled over this today myself when using vmdebootstrap where I ran into the same issue as in [1] As I couldn't reproduce the failure with cdebootstrap or debootstrap from stretch, I ran a git bisect against debootstrap. The first faulty commit is https://salsa.debian.org/installer-team/debootstrap/commit/7e533a672ce413fc591b59e83f7c665c1a2e174e And sure enough, reverting that commit fixed the dangling machine-id mount for me. Henrich, do you want a bug report against debootstrap for that? Regards, Micahel [1] https://bugs.debian.org/899155 -- Why is it that all of the instruments seeking intelligent life in the universe are pointed away from Earth?
Attachment:
signature.asc
Description: OpenPGP digital signature