John,

when Ubuntu shuts down, it will send all processes a TERM signal, i. e., politely asking them to quit. After a certain length of time, it will send any remaining processes a KILL signal, forcing them to quit. If you kill -KILL SurgeMail, that’s an unclean shutdown.

Presumably, SurgeMail is taking a long time to shut down, perhaps ChrisP has some ideas about that, or maybe it just has a lot on its plate? When you run

	# tellmail shutdown

how long does it take until all SurgeMail processes exit?

I don’t use Ubuntu myself, but I the shutdown is controlled by systemd, and I understand that if you set the line

	DefaultTimeoutStopSec= 

in /etc/systemd/system.conf to a higher value (e. g. 240), that will give processes more time to shutdown before they are killed. The default is 90 seconds AFAIK.

Best,
Chris


> Am 09.03.2018 um 04:35 schrieb John Wilkes <john@wilkes.com>:
> 
> I’ve been running 7.3c3-3 on Ubuntu 16.04. When I reboot the system I always see an error message on the SurgeMail status page:
> 
> LAST SHUTDOWN WAS UNCLEAN :-) -- Examine logs crash*.log to establish cause
> 
> The log files don’t give any hints about the problem that I can see.
> 
> I am in the process of moving my server to a new location, and I’ve switched to Centos 7. I see the same error there on a fresh, brand new installation of SurgeMail 7.3c3-3 with no accounts and the default surgemail.ini file.
> 
> I did a manual shutdown with "/etc/init.d/surgemail stop” and it shutdown cleanly.
> 
> Any ideas how to get a clean automatic SurgeMail shutdown when I’m shutting down the system? Is this a known problem?
> 
> -- 
> John Wilkes
> john@wilkes.com