* Russ Allbery: > I would dearly like to stop using sysv init scripts for the trivial cases > as soon as I can, since they just introduce a bunch of possible bugs > without much real benefit. Same here. I haven't examined the situation too closely (maybe a more robust variant of start-stop-daemon could do the trick), but with the current approach, there are many different failure modes, but all of them quite rare and thus difficult to fix.