v11: Possibly handle journald restarts better.

I've encountered two runaway tallow daemons now that seem to
coincide with journald restarts that send it spinning tight
on the `continue` statement and hitting the same _get_data()
error (ENOENT).

I'm unsure if the `break` will fix it, but the `continue`
is definitely broken here. Hopefully the `sd_journal_wait()`
will properly reassess the journal state and notify us of
rotations or other issues.