Make jobs next run number immutable to crash about laminar HOT 4 CLOSED

After I restart laminard after a crash there are no finished jobs, so the new jobs get added to the queue, but never start executing, because of this:

laminar.cpp:582: error: Could not create working directory; rd.string() = /var/lib/laminar/run/fun-prime/1

If I delete the ~laminar/run/1 directory then I am getting this:

laminar.cpp:590: warning: Archive directory already exists; archive.string() = /var/lib/laminar/archive/fun-prime/1

from laminar.

Instead of deleting stale directories or forcing a database entry when a job is started rather than completed, how about I just output a warning and continue if the directory already exists? It is anyway a case that should never happen so I prefer not to put too much logic into it

from laminar.

Or clean up the run dir. Delete it maybe.

from laminar.

New behaviour attempts to delete an existing run dir, but will continue anyway if that fails

from laminar.