BUG in the initscripts -- was udev boot problem

Steve Crosby fost at hotmail.com
Wed Dec 29 04:18:03 PST 2004

"Schafer Frank" <Frank.Schafer at t-systems.cz> wrote in 
news:D37EDD418621814DB470E22D6F531A4F0BEF0E at mefisto.t-systems.cz:

> Hmmm,
> Maybe I'm using SysV tooo long. (About 25 years) All U*XES I know
> UNmount filesystems on system halt or reboot except ''/'', which will be
> mounted ro.
> I didn't study /etc/init.d/rc if it makes some differences calling
> scripts in rc0.d or rc6.d on LFS. Normal (and sufficient) behavior is
> calling all K* scripts with stop and all S* scripts with start.

Your correct - it's normal sysvinit behaviour to run K scripts in rc0.d and 
rc6.d to stop services - thats how the LFS bootscripts *used* to work 
(around LFS5ish)

However, a change was made to the rc script sometime since to include the 

        case $runlevel in
                0|6) $i stop    ;;
                *)   $i start   ;;

and to change the K links to S links. Behavoiour is now to have all S 
scripts in runlevels 0 and 6 run with the "stop" parameter. I haven't yet 
found the changelog describing why this change was made however.

- --
Steve Crosby

More information about the lfs-support mailing list