# /etc/init.d/amavisd status * status: stopped # ps auxw | grep amavisd amavis 25253 0.0 3.4 274340 139436 ? Ss 17:59 0:01 /usr/sbin/amavisd (master) amavis 25264 0.1 3.5 282776 141620 ? S 17:59 0:01 /usr/sbin/amavisd (ch8-avail) amavis 25265 0.1 3.5 284260 144404 ? S 17:59 0:02 /usr/sbin/amavisd (ch11-avail) amavis 25266 0.1 3.4 281864 141080 ? S 17:59 0:01 /usr/sbin/amavisd (ch7-avail) amavis 25267 0.1 3.6 286056 146184 ? S 17:59 0:02 /usr/sbin/amavisd (ch10-avail) amavis 25268 0.0 3.4 282004 141064 ? S 17:59 0:01 /usr/sbin/amavisd (ch7-avail) # /etc/init.d/amavisd restart * Starting amavisd ... * start-stop-daemon: /usr/sbin/amavisd is already running * Failed to start amavisd [ !! ] * ERROR: amavisd failed to start As a quickfix amavisd.initd-r1 can be used.
Ironically, I rewrote the init script for v2.11.1 (in part) to fix the exact opposite issue: bug #646336. The new init script runs amavis in the foreground, so it should be easy to figure out what's going wrong. Amavis no longer creates its own PID file, so there's only one PID file at /run/amavisd.pid containing the PID of the "amavis (master)" process. When you try to stop the service, OpenRC should kill the PID contained in that file, and not succeed until the process is dead. But from your "ps" output, it obviously isn't... Can you tell me what's in that PID file after a fresh start, and what happens during/after you tell OpenRC to stop it? What's in the PID file (and what is the PID of the amavis master process) after the problem appears?
Ping =)
Sorry it took me so long and it's a false alarm. The problem was that I had app-admin/monit with old pidfile location. After adjusting the pidfile to /run/amavisd.pid it works properly.