On Fri, 12 Oct 2018, Will Aoki wrote: > If spamass-milter takes too long to create the socket after daemonizing > itself, the socket file will not exist when the startup script attempts > to change its permissions. MTAs relying on the permission change, such > as Postfix, will then be unable to communicate with the milter. > > This is a very rare condition: I've encountered it only once in years of > operation. It may have involved I/O contention with other VMs. > > The following appeared in the system log. Note that systemd thinks > spamass-milter startup failed, but spamass-milter was actually running.
Yeah, you're absolutely right. The correct solution (for systemd) involves having systemd create the sockets with the right permissions, and then pass them on to spamass-milter but that requires significant changes to spamass-milter that I won't have time to create in the near future. I'll certainly accept patches which do that, though. -- Don Armstrong https://www.donarmstrong.com Of course, there are cases where only a rare individual will have the vision to perceive a system which governs many people's lives; a system which had never before even been recognized as a system; then such people often devote their lives to convincing other people that the system really is there and that it aught to be exited from. -- Douglas R. Hofstadter _Gödel Escher Bach. Eternal Golden Braid_
