Hello, When I stop the mg with "/etc/init.d/milter-greylist stop" command on my Centos system I never got clean greylist.db dump. There is always greylist.db-XXXXXX temporary file left over, and the actual dump is from a previous dump event. It looks like that the mg requires on my system (actually quiet performant) more than the default 3 seconds to finish writing the dump. (The killproc sends shortly after the TERM signal a KILL signal to the process) I wanted to make the delay for the hard stop to be customizable, so please include the patch attached, may be useful for others. I changed the default value of the delay to be 5 seconds, but this can be changed as specifying different value for the hardstopdelay variable in the /etc/sysconfig/milter-greylist file. Bests, Attila
Message
Loosing last dump on RedHat based system when stopping mg
2009-12-11 by attila.bruncsak@itu.int