Your message dated Sat, 29 Dec 2018 18:33:31 +0000
with message-id <[email protected]>
and subject line Re: sysvinit: Coherent symlink numbering policy
has caused the Debian Bug report #153743,
regarding Default rc.d 'K' sequence numbers.
to be marked as done.
This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
Bug report if necessary, and/or fix the problem forthwith.
(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact [email protected]
immediately.)
--
153743: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=153743
Debian Bug Tracking System
Contact [email protected] with problems
--- Begin Message ---
Package: sysvinit
Version: 2.85-7
Amazingly, it appears that the Debian script "update-rc.d" defaults to
creating 'S' and 'K' rc.d symlinks with a sequence number of 20 and 20,
respectively. I'm sure that everyone knows that the generally accepted
way to pair these values is to have the 'K' value be 100 minus the 'S'
value, so that the shutdown order is inverse the startup order.
I understand that (somehow) Debian has survived with this fundamentally
flawed default, and that it is hard to change something that so many
people depend on. However, I think it is worse to not address the
problem.
Naively, I would argue that the defaults should make 'K' sequence numbers
80 by default.
However, I note that on my Debian installs, the sequence numbers between
20 and 80 are virtually un-used. I contend that this is because of the
defaults make in unclear what will be started / stopped in that range.
Also, for simplicity's sake, I think using values of 50 for both 'S' and
'K' sequence number would be more appropriate and less confusing to
novices of the SysV Init system.
If none of these avenues of action seem viable, then I would argue that
the defaults should be explicitly removed. This would break every package
that depends on it, (and it seems like there are many many that do,) but
those package's maintainers would be be forced to evaluate the appropriate
'S' and 'K' values for their services. This would also enlighten them
that the system before may not have been assigning the assumed inverse 'S'
value to the 'K' links.
-- Charles Wilcox
P.S. It is things like this that make people dislike Debian. Heck, I'm
considering using Slackware or Gentoo because the basics of those Linux
distributions are better.
Debian tries to be the one distribution that seems to be more ideal driven
in many areas. I really want to support Debian. But it needs to have the
logistics right, because if the functionality is broken, less people will
use it and they will miss out on the benefits of Debian.
--- End Message ---
--- Begin Message ---
[2002-07-21 15:48] "Peter L Jones" <[email protected]>
> Package: sysvinit
> Version: 2.84-3
> Severity: wishlist
>
>
> Currently the debian-policy section 10.3 gives little or no guidance
> on how scripts should be numbered. I think the following rules could
> (and should) be encapsulated by the update-rc.d routine:
>
> 1) You must populate shutdown (K*) scripts in all directories you don't
> put startup (S*) scripts in.
> -- This ensures clean shutdowns and runlevel changes.
> -- Doesn't apply to S* scripts put in rc[016].d, of course.
>
> 2) The mm parts of you S* and K* scripts must add up to 100 (e.g.
> rc2.d/S81timidity -> rc0.d/K19timidity).
> -- This ensures the correct order of shutdown.
>
> 3) If you put an S* script in rcS.d, there should be an entry in rc[06].d
> (named S*)
> -- The mm parts must add up to 100.
> -- This ensures the correct order of shutdown.
I pretty sure, with introduction of dependency-based boot and insserv(8)
this bug could be closed.
--- End Message ---
