New ZeroConf Spec

Trent Lloyd lathiat at bur.st
Wed Jul 5 08:03:20 BST 2006


On Wed, Jul 05, 2006 at 08:43:02AM +1000, Matthew Palmer wrote:
> On Tue, Jul 04, 2006 at 02:48:22PM +0800, Trent Lloyd wrote:
> > On Tue, Jul 04, 2006 at 04:03:05PM +1000, Matthew Palmer wrote:
> > > On Tue, Jul 04, 2006 at 11:35:59AM +0800, Trent Lloyd wrote:
> > > > I also feel that removing avahi in synaptic is not the way to go about
> > > > disabling avahi, that can be done simply by stopping and preventing the
> > > > daemon from starting on boot with the /etc/default infrastructure that
> > > > is now in debian.
> > > 
> > > Using /etc/default is Just Plain Wrong, because it prevents you from
> > > starting the service at any other time, too.  The *correct* way to stop a
> > > service from starting at boot is to modify the symlinks in /etc/rcN.d. 
> > > sysv-rc-conf is available (amongst other tools) to manipulate the symlinks
> > > on your behalf if you're not comfortable doing it by hand.  update-rc.d is
> > > not recommended, due to being a mighty poor UI for the command line.
> > 
> > This is true and has often annoyed me, however, can sysv-rc-conf restore
> > the links back to their original levels?
> 
> When you say "level", I presume you're talking about ordering -- so 'level'
> within a runlevel, as opposed to which runlevels the service starts in.
> 
> It's not *quite* optimal in it's handling of ordering -- when you switch a
> service from S to K, or vice-versa, it leaves the ordering number the same,
> so, for example (an example I just tried on my laptop), S91apache becomes
> K91apache when I uncheck the box in sysv-rc-conf, and then K91apache becomes
> S91apache when I check it again.
> 
> The correct behaviour, of course, is to make the ordering inverse, so the
> ordering number should be 100-<current> when switching -- so S91apache
> becomes K09apache.  It's not a massive problem, though, when your primary
> purpose is "don't start at boot", and not "make sure everything works right
> when I transition between runlevels", because runlevel transition (apart
> from S->N and N->[06]) rarely happens on modern systems.
> 
> > I noticed update-rc.d didnt' seem to have any idea what they were to
> > start to (run order etc)
> 
> Yep, update-rc.d doesn't have any knowledge of that, nor is it intended to
> -- it's sole purpose in life is to run in a maintainer script and set the
> maintainer's idea of default run levels.  Do one job and do it well, and all
> that.

Which begs the question is there a tool that does know?

Trent

> 
> - Matt
> 
> -- 
> I'm not sure which upsets me more: that people are so unwilling to accept
> responsibility for their own actions, or that they are so eager to regulate
> everyone else's.
> 
> -- 
> ubuntu-devel mailing list
> ubuntu-devel at lists.ubuntu.com
> https://lists.ubuntu.com/mailman/listinfo/ubuntu-devel

-- 
Trent Lloyd <lathiat at bur.st>
Bur.st Networking Inc.



More information about the ubuntu-devel mailing list