[Pkg-pdns-maintainers] Bug#768078: 30 second delay on shutdown

[Serge Belyshev]

> I disagree. pdns-recursor cleanly exits on the first HUP, and the
> TERM/KILL will never be triggered.
>
> If pdns-recursor doesn't immediately exit, something's fishy.

No objection here.

> In other news, the systemd unit file for pdns-recursor doesn't do
> anything fancy, and a stop would run the regular systemd exit
> behavior (which means way faster timeouts even if the daemon hangs
> or something).

OTOH, sysvinit implements exactly the same feature by itself: when
changing runlevels, it sends TERM and then KILL signals to processes
that are stuck.