Le vendredi 5 décembre 2008 12:35:44 Craig Small, vous avez écrit : > On Fri, Dec 05, 2008 at 12:29:43PM +0100, martin f krafft wrote: > > also sprach Craig Small <[EMAIL PROTECTED]> [2008.12.05.1211 +0100]: > > > There is no sensible fix for this. The problem is that any time is the > > > wrong time for the init script to run, for someone. > > > > Do you have specific examples that would break if procps ran at S40? > > Shifting it to 40 would probably fix your ipv6 related bug, I guess. > > There are some settings that need to be set before the interfaces come > online, I'm not sure of the specifics but I do know these keys are out > there and they do need to be set before networking starts. > > Ideally there probably should be some sort of runlevel type sysctl > setup, that's going to be fiddly and not trivial to write, but not > impossible either. > > My worry is that we move it, you're happy, people who have similiar > problems to you are happy but then suddenly another group of people > spring up with "hey, why did you move sysctl, you broke my system now" > > - Craig
As ipv6 is a release goal, my guess is that a solution has to be found.
Then, why not getting opininons by consulting -devel about
* what would happen if procps was moved
from S30
Required-Start: mountkernfs
to S45
Required-Start: mountkernfs $network
(actually, AFAII, the number is somewhat "automagic", what matters is
the "Required-Start")
* other solution ideas
?
Regards, OdyX
--
Didier Raboud, proud Debian user.
CH-1802 Corseaux
[EMAIL PROTECTED]
signature.asc
Description: This is a digitally signed message part.
