Here is the patch I propose to fix this. The patch adds a check in do_start() which scans the contents of the PID file (if it exists), and checks that the PID belongs to a running instance of system-tools-backends. If it does, then it exits gracefully.
This patch is already in Jaunty (and Debian), so I just backported it. It could actually be fixed with a much more simple patch, just by passing "--oknodo" to start-stop-daemon. However, Debian did it like this and I assume there is probably a good reason for it.
Here is the patch I propose to fix this. The patch adds a check in do_start() which scans the contents of the PID file (if it exists), and checks that the PID belongs to a running instance of system- tools-backends. If it does, then it exits gracefully.
This patch is already in Jaunty (and Debian), so I just backported it. It could actually be fixed with a much more simple patch, just by passing "--oknodo" to start-stop-daemon. However, Debian did it like this and I assume there is probably a good reason for it.