systemd timeout on startup after upgrade

[Steve Litt]

Felix Zielcke said on Sat, 17 Apr 2021 19:37:30 +0200

>Hello *,
>
>I have upgraded today a Debian buster (stable) system to bullseye
>(still testing). After upgrade I did a complete reboot of the VM.
>
>Dovecot version changed from 1:2.3.4.1-5+deb10u6 to  1:2.3.13+dfsg1-1
>
>`systemctl start dovecot` now complains
>Job for dovecot.service failed because a timeout was exceeded.
>
>As a workaround I set "TimeoutStartSec=infinity" in dovecot.service.
>But `systemctl status` now says:

Systemd's got problems. I use runit to start dovecot. 

Without dumping systemd, you can start
the runit process supervisor from systemd, and then start any other
daemons from runit. I used to use this method (with daemontools
instead of the very similar runit) to avoid starting from sysvinit. I
don't use systemd, but if I did, I'd use it as little as possible,
because it's a very complicated, very ever-scope-expanding moving
target.

I don't even call the above described method as a workaround, because I
consider systemd the root cause.

SteveT

Steve Litt 
Spring 2021 featured book: Troubleshooting Techniques of the Successful
Technologist http://www.troubleshooters.com/techniques