FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Debian > Debian Development

 
 
LinkBack Thread Tools
 
Old 01-04-2008, 05:43 PM
"Joe Smith"
 
Default Faster shutdown and the ubuntu "multiuser" update-rc.dextention

"Josselin Mouette" <joss@debian.org> wrote in message
news:1199470514.4357.107.camel@shizuru...

Ubuntu discovered this a while back, and introduced a method to avoid
calling stop scripts in runlevel 0 and 6. It is the "multiuser"
extension to update-rc.d, and in Ubuntu packages are changed to calls
dh_installinit with '-- multiuser' as an argument to enable it. This
add the "multiuser" argument (instead of to the "default" argument) to
update-rc.d, which go on and set up the boot sequence without
references to the script in runlevel 0 and 6. This can be done
without such extention, and how is the topic of the rest of my email.


Frankly I couldn’t imagine a worse idea to fix this problem.

Many daemons will corrupt their state if they aren’t killed cleanly.
Leaving them a grace time is actually worse than simply cutting the
power, because you can be sure the daemon is actually writing some data
at the moment you kill it.


Really. First AIUI most of the candidates for this already do something
similar
in their shutdown scripts. The sigkill ensures that a program which hanged
in the sigterm handler
gets terminated. Further, it ensures that any service that chose to ignore
the sigterm signal for some

odd reason gets killed as well.



The most funny thing comes from daemons which depend on each other. You
will easily hit cases where a daemon will not shut down properly because
it cannot find the one that depends on it, and in the end increase the
shutdown time instead of shortening it.


Those cases are also not candidates for this.


Let’s just switch to a parallel init system, and the Ubuntu wannabees
will win their precious few seconds without risking to introduce data
corruption bugs on production systems.


While those with dependecies and complex clean shutdown procedures would
benefit from such a system, the remaining
cases would still be sped up slightly more by letting the later script take
care of them.



On a somewhat related note, it has just occured to me that the standard
GNU/Linux system design (and as far as i know, the standard UNIX design)
does not really have a concept of shutdown priority. I know that Microsoft
Windows does have some meathod of indicating a higher than normal priority
shutdown (useful for things like autoshutdown when the UPS indicates that it
is running out of power). Should a similar sytem of some kind be implemented
in Debian GNU/Linux?




--
To UNSUBSCRIBE, email to debian-devel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
 
Old 01-04-2008, 06:41 PM
Pierre Habouzit
 
Default Faster shutdown and the ubuntu "multiuser" update-rc.dextention

On Fri, Jan 04, 2008 at 06:43:54PM +0000, Joe Smith wrote:
>
> "Josselin Mouette" <joss@debian.org> wrote in message
> news:1199470514.4357.107.camel@shizuru...
> >>Ubuntu discovered this a while back, and introduced a method to avoid
> >>calling stop scripts in runlevel 0 and 6. It is the "multiuser"
> >>extension to update-rc.d, and in Ubuntu packages are changed to calls
> >>dh_installinit with '-- multiuser' as an argument to enable it. This
> >>add the "multiuser" argument (instead of to the "default" argument) to
> >>update-rc.d, which go on and set up the boot sequence without
> >>references to the script in runlevel 0 and 6. This can be done
> >>without such extention, and how is the topic of the rest of my email.
> >
> >Frankly I couldn’t imagine a worse idea to fix this problem.
> >
> >Many daemons will corrupt their state if they aren’t killed cleanly.
> >Leaving them a grace time is actually worse than simply cutting the
> >power, because you can be sure the daemon is actually writing some data
> >at the moment you kill it.
>
> Really. First AIUI most of the candidates for this already do something

Most is not all. Some (many ?) daemons have proper shutdown sequence
based on control sockets e.g. Or on SIG*TERM*s and not KILL.
--
·O· Pierre Habouzit
··O madcoder@debian.org
OOO http://www.madism.org
 

Thread Tools




All times are GMT. The time now is 08:34 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright 2007 - 2008, www.linux-archive.org