Comment # 43 on bug 1094762 from andreas bittner

If you give me detailed instructions on how to set loglevels that will help
you, I have still that one unupgraded machine here, I need to do something
about it.

I dont know though if the machine will be affected though.

What comes to my mind maybe is, is it possible that non-rebooted machines that
e.g. had updates on 42.3 level (systemd and other stuff lately, not yet
replaced and activated all files in use, e.g. non rebooted yet) and then going
for a zypper dup with repositories swapped over to 15.0 URL addresses

maybe can cause this bug? just a gut feeling, maybe I am totally wrong. So
pending systemd (and the processes you speak about) updates and then in
addition to that zypper dup to 15.0 and then maybe restarts of services fail
differently than on a cleanly booted 42.3 system with nothing pending?