So, it seems that this problem exists because systemd is trying to mount /var/log (which is already mounted at that point, and this fails), because of this utmp can't be updated which leads to the hang, a log:

    Mounting /var/log...
    var-log.mount: Directory /var/log to mount over is not empty, mounting anyway.
    var-log.mount failed to run 'mount' task: No such file or directory
    var-log.mount changed dead -> failed
    Job var-log.mount/start finished, result=failed
    [FAILED] Failed to mount /var/log.
    See 'systemctl status var-log.mount' for details.
    Job systemd-update-utmp-runlevel.service/start finished, result=dependency
    [DEPEND] Dependency failed for Update UTMP about System Runlevel Changes.
    Job systemd-update-utmp-shutdown.service/start finished, result=dependency
    [DEPEND] Dependency failed for Update UTMP about System Shutdown.
    Unit var-log.mount entered failed state

Lennart,

a bit of background: The servic ewhich is running into this bug is an auto-installer which modifies /etc/fstab (so fstab at the beginning is different to the fstab at the end).
Initially /var/log is part of the fs root /.
During the installation an fstab entry is added which maps /dev/mapper/HostVG-Logging to /var/log.

IIUIC systemd now tries to do a mount according to that fstab entry, but fails (I dont know why it fails, I don't see a reason for it in the logs).

The question is, how can this be solved cleanly? And have you got an idea about the failing mount?

A dirty workaround:
http://gerrit.ovirt.org/#/c/12935/

Hmm, the "No such file or directory" here happens when the fstab first had an entry for the unit, and by the time the unit is about to to be activated the entry is gone.

Is that what happens here?

Seeing a similar problem again, but without the messages given in comment 2.