Re: Strang shutdown mechanism with Debian
In article <[🔎] 19980517090530.A10783@kuolema.Infodrom.North.DE>,
Martin Schulze <joey@infodrom.north.de> wrote:
>This is not a clean method. I'm dissappointed.
Yes, well if you write a system from scratch you can do it cleanly. For
now we had to work whatever was available and would break too much.
>What do you mean by "most of the levels in Kxx were already taken"?
>
>Only urandom contains diversion between "start" and "stop".
>(mdutils.sh is non-existant, it would behave different, too.)
Those two, yes. There might be more of which we know nothing yet but
it's good to be prepared for the future.
>For me it would make more sense moving the S*scripts that
>need to be called with "stop" to K92..K98. I don't see that
>Knn is "full".
And what if you want to insert a new script somewhere in between? That
would not be possible anymore. The Kxx and Sxx entries need to be sparse
so that you can move things around and insert new links "in between".
I though about moving the Kxx links, but that would mean filing bugreports
against all packages that install links in /etc/rc?.d and fixing them. That
would probably take _months_ to get sorted out and will no doubt break a
lot of systems.
>Apart from that, why is sendsigs called _before_ all scripts
>got the chance to terminate their services? This broke the
>shutdown mechanism with recent file-rc installations.
The time at which sendsigs (killall -TERM; killall -KILL) got called
hasn't been changed since Debian 1.0.0. The exact same order was
used in the old monolithic halt and reboot scripts...
Mike.
--
Miquel van Smoorenburg | Our vision is to speed up time,
miquels@cistron.nl | eventually eliminating it.
--
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: