View previous topic :: View next topic |
Author |
Message |
tilt n00b
Joined: 16 Nov 2003 Posts: 36 Location: Stuttgart
|
Posted: Fri Nov 12, 2004 8:26 am Post subject: hangs on remounting remaining filesystems readonly |
|
|
This problem has been bugging me for some time: When I shut down or reboot, the system hangs after "remounting remaining filesystems readonly". I've searched the forums, and this is not the usual APM problem. There is never an [OK] after that message. I've put some debug output in /etc/init.d/halt.sh to see what's going on. First, this is the script's problem part:
Code: | ebegin "Remounting remaining filesystems readonly"
# Get better results with a sync and sleep
sync; sync
sleep 1
if ! mount_readonly
then
killall5 -9 &>/dev/null
sync; sync <----------------------- locks up here
sleep 1
if ! mount_readonly
then
eend 1
sync; sync
[ -f /etc/killpower ] && ups_kill_power
/sbin/sulogin -t 10 /dev/console
else
eend 0
fi
else
eend 0
fi
|
I've found out that "mount_readonly" tries to unmount /var (which is on its own partition) first, succeeds, then tries to unmount / twice! Both times, mount returns a value of 32 (according to its man page that's a mount failure). The script then enters the if branch and hangs on the "sync; sync" line. The sleep command never gets executed.
The last time everything went fine was with gentoo-dev-sources 2.6.7. I've updated to 2.6.8 and the bug appeared, and now on 2.6.9 it's still there.
edit: shortened title |
|
Back to top |
|
|
tilt n00b
Joined: 16 Nov 2003 Posts: 36 Location: Stuttgart
|
Posted: Sat Nov 13, 2004 11:34 am Post subject: more details |
|
|
Hrm... seems I'm the only one with that problem
I've figured out some more details, but don't have a solution yet. First, the reason given why / fails to remount is:
The halt.sh script gets the partitions it's supposed to remount from /proc/mounts. This is how my /proc/mounts looks at that point:
Code: | rootfs / rootfs rw 0 0
none /sys sysfs rw 0 0
none /proc/bus/usb usbfs rw 0 0
none /proc proc rw,nodiratime 0 0
none /dev/pts devpts rw 0 0
none /dev ramfs rw 0 0
/dev/root / reiserfs rw,noatime 0 0
/dev/hdb8 /var reiserfs ro 0 0 |
The first thing that strikes me are "rootfs" and "/dev/root". I've never seen those before. Seems like it has something to do with udev? (I'm running a udev-only system, latest udev ebuild installed).
Clues, anyone? |
|
Back to top |
|
|
tilt n00b
Joined: 16 Nov 2003 Posts: 36 Location: Stuttgart
|
Posted: Sun Nov 14, 2004 3:54 pm Post subject: |
|
|
Seems like I've found the culprit. It isn't a kernel issue after all.
The problem won't appear if I don't start X (rc-config del xdm default).
If I start X, however, remounting the filesystem fails. A quick call of "ps auxf" revealed the (possible) culprit: bonobo-activation-server and evolution-data-server-1.0 are still running and are probably blocking the read-only remount.
Those two get killed by the "killall5" command in halt.sh, but then the system locks up at the "sync; sync" line (see 1st post).
So... still no solution to my problem, but at least some info to file a bug report. Can anyone of the mods please move this thread out of "Kernel & Hardware"? |
|
Back to top |
|
|
tilt n00b
Joined: 16 Nov 2003 Posts: 36 Location: Stuttgart
|
Posted: Sun Nov 21, 2004 3:51 pm Post subject: |
|
|
Just to tie up this loose end: Bug report is here, and a workaround is to put
Code: | killall evolution-data-server |
into /etc/conf.d/local.stop |
|
Back to top |
|
|
rhill Retired Dev
Joined: 22 Oct 2004 Posts: 1629 Location: sk.ca
|
Posted: Mon Nov 22, 2004 1:21 am Post subject: |
|
|
glad we could help.
|
|
Back to top |
|
|
|