how to debug systemd unit problems?

[Corey Hickey <bugfood-ml@fatooh.org>]

Hi,

With a recent upgrade, my sid system ended up with systemd-sysv, so I
rebooted it and worked through the problems. One problem I had a hard
time figuring out--I kind of guessed at how to fix it. I'd really like
to know the right way, though, so as to be able to troubleshoot problems
like that in the future.

With a regular init script, I am used to doing the following:
bash -x /etc/init.d/foo

How do I do something similar with a systemd unit? I can't figure out
how to find any indication of what programs are actually being run in
order to fulfill a unit.

The unit in question was dev-mapper-common.device, and all journalctl
ever said was that it timed out. I'd paste all the information here, but
I can't figure out how to get information from prior to the current boot
either. Fixing the problem entailed finding an entry in /etc/fstab that
referenced a device "/dev/mapper/foo" that doesn't exist unless I create
it manually; I added the "noauto" parameter--I guess this failed in a
non-fatal fashion before.

Thanks,
Corey