[systemd-devel] [PATCH] core/mount: only add dependencies to active units for dynamic mounts [was: Re: unit: When stopping due to BindsTo=, log which unit caused it]

[Andrei Borzenkov]

В Thu, 26 Feb 2015 21:59:24 +0100
Martin Pitt <martin.pitt at ubuntu.com> пишет:

> Hello again,
> 
> Martin Pitt [2015-02-26 21:40 +0100]:
> > Ah, then I misunderstood the problem, and the patch will most probably
> > not work.
> 
> Sorry, that was too fast. It actually should work. On second thought,
> it doesn't matter if there's systemd in the initrd or not; if it gets
> stopped before pivoting root, and a new one gets started in the new
> root, that's exactly the same behavior as having no systemd in the
> initrd.
> 

No, it's not the same. systemd is expected to save and reload its state
across switching to new root.

> > Most probably your initramfs has systemd, but it gets stopped and the
> > "real" root file system gets another systemd instance after pivoting.
> > The mount is still in /proc/mounts, but due to the restarting systemd
> > all of its units go away. Is that the case?
> 

"all units go away" would be a bug. It could be another case of
https://www.mail-archive.com/systemd-devel@lists.freedesktop.org/msg28398.html
(systemd attempting to react on incomplete state).