In the stage 1 script you listed, I can't find a mount instruction for /run. If you are running that stage 1 script verbatim, you are making /run/s6
> on your root filesystem, not on a tmpfs. If you are not running that script
> verbatim, please provide the *exact* script you are using, else I can't
> help you.
Sorry, the initramfs do the stuff like this :
mount -t tmpfs run /run -o nosuid,nodev,mode=0755
The init on initramfs swicth the root with this classic command
exec env -i "TERM=$TERM" /usr/bin/switch_root /new_root $init "$_at_"
The append line in syslinux.cfg is :
APPEND root=/dev/sdb4 vga=895 rw init=/etc/s6/init
so normaly when i enter in the init script, the / is mounted rw.
> the *exact* script you are using
This is the exact script.
This is the script for stage 2
if -nt {
if { s6-echo "starting stage2" }
# start service
if { s6-rc-init -c /etc/s6/stage2serv /run/s6-run/service }
if { s6-rc -u change all }
s6-echo "stage2 completed."
}
redirfd -w 1 /dev/console
s6-echo "Something wrong on stage2, log on tty3 to debug it"
> This is expected, and you don't have to start it manually either. If
> it's not brought up, it's simply that you don't need it. If s6-rc runs a
> service that depends on it, it will bring it up automatically.
> (For now, s6rc-fdholder is only needed when you have longrun pipelines.
> If you don't have pipelines, nothing depends on it and it can remain down.)
OK, thank
--
Eric Vidal <eric_at_obarun.org>
Received on Fri May 13 2016 - 14:24:40 UTC