System Halt and udev Errors

BlueFusion · Guru Joined: 08 Mar 2006 Posts: 371

Hello Gentoo users,

I have searched up and down on the net for a problem I've been having for a few days now. The problem arose when I compiled the new 2.6.16-gentoo-r3 kernel. Everything has been running great after customizing it like I did the previous kernel, except when I halt the system (not reboot).

The regular service stop procedures run through normally, right to the point of where it says "System Halt." After a pause of 1 to 2 seconds, errors start going down the screen. The errors are too fast to read, thus I can't repeat them here in full. The errors go on until I manual press the power button to the system.

From what I can see in the never-ending screen of errors is that there are three lines that repeat over and over. All of which begin with "udev event" and end with "read-only filesystem."

As I mentioned before, this only occurs on system halt and not on reboot.

For a control test, I booted back up with my 2.6.15-gentoo-r1 kernel to try to reproduce the same things, but no errors. I figured that somehow, someway, it was because I didn't enable Power Management into the kernel (although that's the next thing to get setup on my system). I recompiled the kernel with Power Management support, still not selecting the ACPI or APM features (same as the 2.6.15-gentoo-r1 settings I was using). This, still, provided no new luck.

Does anyone have any ideas as to how I can fix this error so the system will shut down properly or to even find out what the error messages are so I can dig deeper into this?
_________________
i7-940 2.93Ghz | ASUS P6T Deluxe (v.1) | 24GB Triple Channel RAM | nVidia GTX660
4x 4TB Seagate NAS HDD (Btrfs raid5) | 2x 120GB Samsung 850 EVO SSD (Btrfs raid1)

jossan · Posted: Fri Apr 28, 2006 5:40 pm Post subject: Re: System Halt and udev Errors

BlueFusion · Guru Joined: 08 Mar 2006 Posts: 371

After getting partyway through `emerge -e system` (see threads on that problem here and here), I shutdown the system to give the A/C a chance to cool my room again :roll:

. Anyway, this time, when it shutdown, it just hadd the following after the end:

wjb · Posted: Mon May 01, 2006 11:32 am Post subject:

I had the problem with the messages at shutdown from somwhere around middle of March after I moved to gentoo-sources-2.6.15-r1

It seemed to be trying to delete something from /dev, but after the point of reporting the remaining file systems being made read only. I could never read the messages ("...failed to delete ...") before the power down.

My problem seems to have [mostly] gone away in the last couple of weeks - possibly baselayout (which changed about then) but I'm not really sure. Its usually ok, but sometime it still goes wrong.

jossan · Posted: Tue May 02, 2006 5:43 pm Post subject:

Still broken with gentoo-sources-2.6.16-r5 here. :-(

_________________
From the moon we'r all al the same.

wjb · Posted: Thu May 04, 2006 7:24 pm Post subject:

This is from a pic I took of the screen as the system shut down. It was a bit blurry so some of its guesswork - the digits mostly, but also the hiddev0 stuff.

BlueFusion · Guru Joined: 08 Mar 2006 Posts: 371

That looks like what I have, too.

Now running 2.6.16-gentoo-r6...same thing.
_________________
i7-940 2.93Ghz | ASUS P6T Deluxe (v.1) | 24GB Triple Channel RAM | nVidia GTX660
4x 4TB Seagate NAS HDD (Btrfs raid5) | 2x 120GB Samsung 850 EVO SSD (Btrfs raid1)

jossan · Posted: Thu May 04, 2006 8:53 pm Post subject:

It would be great to letting us know what extra information is usefull to help fixing this problem. :roll:

Perhaps it'is a problem in: /etc/group

Jimmy Jazz · Guru Joined: 04 Oct 2004 Posts: 333 Location: Strasbourg

Hello BlueFusion,

I get a similar problem with udev-090 but during the boot process. udevtrigger didn't work well on my box.
You probably use RC_DEVICE_TARBALL that help you to boot. In my case i were not able to boot at all (i don't use device tarball).
Also your problem could be related with udev-start.sh.
Have a look at my bug report #13462.
https://bugs.gentoo.org/show_bug.cgi?id=132462

Hope that will help you

Jj

BlueFusion · Guru Joined: 08 Mar 2006 Posts: 371

I'm using udev-087. It boots up without any errors and reboots just fine. It's just poweing down that fails. When the screen says "Power Off." it just starts with the udev errors or it just hangs there.

Here's my systm info:

emerge --info

Jimmy Jazz · Guru Joined: 04 Oct 2004 Posts: 333 Location: Strasbourg

@BlueFusion

if you set RC_DEVICE_TARBALL to yes, you are sometimes able to boot even when udev doesn't create /dev files.
Perhaps are you lucky, but that doesn't mean udev-start.sh works as it should. So during power off udev-stop.sh returns you an error.

The problem seems to be corrected with udev-090

PS: you can find udev-{start,stop}.sh in /lib/rcscripts/addons

Jj

BlueFusion · Guru Joined: 08 Mar 2006 Posts: 371

I changed the following in /etc/conf.rc:

RC_DEVICES="auto" to RC_DEVICES="udev"
RC_DEVICE_TARBALL="yes" to RC_DEVICE_TARBALL="no"
RC_PARALLEL_STARTUP="no" to RC_PARALLEL_STARTUP="yes"

Thus far, same errors on system halt.

You think updating to udev-090 will fix this?
_________________
i7-940 2.93Ghz | ASUS P6T Deluxe (v.1) | 24GB Triple Channel RAM | nVidia GTX660
4x 4TB Seagate NAS HDD (Btrfs raid5) | 2x 120GB Samsung 850 EVO SSD (Btrfs raid1)

sunckell · n00b Joined: 20 Apr 2005 Posts: 19

I am getting these errors as well. My computer never shuts down. It just failes on the removal of the dev files.

For some reason it seems that "/lib/rcscripts/addons udev-stop.sh" is called. The last executed line in that file is:

BlueFusion · Guru Joined: 08 Mar 2006 Posts: 371

After re-compiling my kernel with cpufreqd and ACPI support, it works now. Makes sense it needs power management to tell the system to halt, but I could have sworn I already tried that

.
_________________
i7-940 2.93Ghz | ASUS P6T Deluxe (v.1) | 24GB Triple Channel RAM | nVidia GTX660
4x 4TB Seagate NAS HDD (Btrfs raid5) | 2x 120GB Samsung 850 EVO SSD (Btrfs raid1)

phsdv · Guru Joined: 13 Mar 2005 Posts: 372 Location: Europe

Same (or similar) issue here. I am using udev-087. Just before power-off I get the following messages:

ReD-BaRoN · Apprentice Joined: 06 Feb 2004 Posts: 208

So is there a bug filed for this specific issue? There should never be a time, ACPI complied or not, where the is a runaway loop, not allowing the computer to shut down.

Thanks!

linoleum · Tux's lil' helper Joined: 15 Nov 2004 Posts: 129

I'm having the same problem with udev-087 and kernel 2.6.16-r7

Any solutions?

ReD-BaRoN · Apprentice Joined: 06 Feb 2004 Posts: 208

Try enabling ACPI in the kernel and see what happens. It seemed to fix it for me, but then I think it's back. I haven't had time to test it throughly yet.

linoleum · Tux's lil' helper Joined: 15 Nov 2004 Posts: 129

ACPI is enabled but the problem is still there.

ReD-BaRoN · Apprentice Joined: 06 Feb 2004 Posts: 208

How about APM?

linoleum · Tux's lil' helper Joined: 15 Nov 2004 Posts: 129

APM also enabled.

ReD-BaRoN · Apprentice Joined: 06 Feb 2004 Posts: 208

Then I'm not sure sorry

.

mjwelsh · n00b Joined: 05 Jun 2006 Posts: 2

I re-emerged udev-090 then did
rc-update del coldplug then did
rm -f /etc/init.d/coldplug

which fixed it - system now shuts down normally

ReD-BaRoN · Apprentice Joined: 06 Feb 2004 Posts: 208

What does coldplug do?

mjwelsh · n00b Joined: 05 Jun 2006 Posts: 2

coldplug detects your hardware and loads modules for it if needed - udev does the detection itself so it doesn't need coldplug anymore