Summary: | Cannot halt the system with baselayout-1.12.7-r4 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Fab <netbox253> |
Component: | [OLD] baselayout | Assignee: | Gentoo's Team for Core System packages <base-system> |
Status: | VERIFIED NEEDINFO | ||
Severity: | normal | ||
Priority: | High | ||
Version: | 2006.1 | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- |
Description
Fab
2006-12-18 02:28:42 UTC
Have a long hard look at bug #156766 Do as much debugging as you can in /etc/init.d/halt.sh to find out what your error is so I can fix it. Ok, I will try to tweak this script to see what's going on, but I need time to do it. I know that /etc/init.d/shutdown is called at the end. --- echo "going to load /etc/init.d/$1" # Load the final script depending on how we are called [[ -e /etc/init.d/"$1".sh ]] && source /etc/init.d/"$1".sh --- Do you know if there's a way to relaunch the system without really rebooting ? Currently I'm using the magic sys keys to do it, but it take time to reboot just to do a test. Thanks. Does /etc/init.d/shutdown.sh exist? To save time you can use kexec (In reply to comment #3) > Does /etc/init.d/shutdown.sh exist? > # ls /etc/init.d/ | grep shut # Strange... It does not, I don't know why. I will try to re-emerge sysvinit. I found another problem during the "unmounting filesystems" step : my partitions are unmounted in this order : * Unmounting filesystems /var/tmp/ccache /var/portage /var/ /usr /tmp /temporaire /opt /home /distpack /data The problem is the following : accordingly to my fstab, /opt is a bind to /usr/opt : --- $ mount ... /dev/hdd1 on /usr type ext3 (rw,noatime) /usr/opt on /opt type none (rw,bind) ... --- It means that /opt should be unmounted before /usr. Currently it doesn't, and the script is trying to force the unmount of /usr because it fails before. Re-emerging sys-apps/sysvinit-2.86-r6 solved the shutdown problem. I don't know why the shutdown.sh script has disappeared. What do you think about the bind's unmount problem ? Do you think you will fix it ? Anyway, thanks for you help. Shouldn't be a problem really as the actual mount point will be cleanly unmounted later. We can't actually test if it's a mountpoint or not as the kernel does not tell us this in /proc/mounts. Otherwise we would unmount the binds first. Yes. I made some tests : I tried to unmount binds before regulars mount points : it changed nothing for /usr : the script still failed on /usr and tries to force to umount it. Probably a process which is used at this moment. I close this bug. Thanks. |