with proftpd-1.2.9_rc1, /etc/init.d/proftpd restart sometimes fails starting it again because the old proftpd didn't yet shutdown. Adding a -R 20 or something to the start-stop-daemon line should do fine for most users.
Can you provide a set of circumstances that cause this problem? If/when I can reproduce it, I'll absolutely add an appropriate delay to the script.
Just tried and failed... background: I was restarting proftpd from a cron job every night (for log rotation). Sometimes it worked, sometimes it failed. It seems there are circumstances where the daemon does not exit fast enough, but I am unable to give better data for testing. Sorry. One possible scenario: The daemon is partially swapped out, and it takes some time to bring the pages in which are needed for signal handling when it's killed.
not only proftpd does this this is a really heavy issue for producing servers, since logfile rotation is really required
Hey... I noticed you just marked 1.2.9 stable for x86
adding --retry 20 to the start-stop-daemon line is a very wise idea.
Added in CVS.