Boot failure

Bug #701060 reported by peterzay
8
This bug affects 1 person
Affects Status Importance Assigned to Milestone
plymouth (Ubuntu)
Invalid
Low
Unassigned

Bug Description

Occasionally, 10.04 won't boot at all or hesitate. CTRL-ALT-DEL may help, but sometimes even that won't work. I boot once daily. The problem happens 1-2 times per week. Should I upgrade to 10.10 to solve this?

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: gnome (not installed)
ProcVersionSignature: Ubuntu 2.6.32-27.49-generic 2.6.32.26+drm33.12
Uname: Linux 2.6.32-27-generic i686
Architecture: i386
Date: Mon Jan 10 10:13:31 2011
InstallationMedia: Ubuntu 10.04 LTS "Lucid Lynx" - Release i386 (20100429)
ProcEnviron:
 LANG=en_US.UTF-8
 SHELL=/bin/bash
SourcePackage: meta-gnome2
---
Architecture: i386
DefaultPlymouth: /lib/plymouth/themes/ubuntu-logo/ubuntu-logo.plymouth
DistroRelease: Ubuntu 10.04
InstallationMedia: Ubuntu 10.04 LTS "Lucid Lynx" - Release i386 (20100429)
MachineType: LENOVO 2746CTO
Package: plymouth 0.8.2-2ubuntu2.2
PackageArchitecture: i386
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-31-generic root=UUID=17572f50-ecf1-466d-a20e-0ded10d25e56 ro
ProcEnviron:
 LANG=en_US.UTF-8
 SHELL=/bin/bash
ProcFB: 0 inteldrmfb
ProcVersionSignature: Ubuntu 2.6.32-31.61-generic 2.6.32.32+drm33.14
Tags: lucid
TextPlymouth: /lib/plymouth/themes/ubuntu-text/ubuntu-text.plymouth
Uname: Linux 2.6.32-31-generic i686
UserGroups: adm audio cdrom dialout dip fax floppy fuse lpadmin netdev plugdev sambashare tape vboxusers video
dmi.bios.date: 05/29/2009
dmi.bios.vendor: LENOVO
dmi.bios.version: 6AET58WW
dmi.board.asset.tag: ATN12345678901234567
dmi.board.name: 2746CTO
dmi.board.vendor: LENOVO
dmi.board.version: LENOVO 6AET58WW
dmi.chassis.type: 10
dmi.chassis.vendor: LENOVO
dmi.chassis.version: LENOVO 6AET58WW
dmi.modalias: dmi:bvnLENOVO:bvr6AET58WW:bd05/29/2009:svnLENOVO:pn2746CTO:pvrThinkPadSL500:rvnLENOVO:rn2746CTO:rvrLENOVO6AET58WW:cvnLENOVO:ct10:cvrLENOVO6AET58WW:
dmi.product.name: 2746CTO
dmi.product.version: ThinkPad SL500
dmi.sys.vendor: LENOVO

Revision history for this message
kurt belgrave (trinikrono) wrote :

Hello there the first thing would really be to check it in natty.
Also if you can take a screenshot so we would have a better idea of the problem you are seeing
Cheers

Changed in ubuntu:
status: New → Incomplete
Revision history for this message
peterzay (peterzay) wrote :

The screen is usually completely black, nothing to see. Occasionally, you may get UBUNTU with the luminous dots underneath.

Revision history for this message
peterzay (peterzay) wrote :

I have just experienced this same intermittent bug in Ubuntu 10.10 on my test machine. Sorry, I don't have another machine to test 11.04.

Revision history for this message
kurt belgrave (trinikrono) wrote :

Can you follow the steps outlined at
https://wiki.ubuntu.com/DebuggingKernelBoot#Boot Options
i will leave the status as incomplete until you respond

Cheers

Revision history for this message
peterzay (peterzay) wrote :

I boot with the Shift (either one, tried both) key held down during BIOS but get no Grub menu. In fact, the PC boots as per normal screens.

Revision history for this message
kurt belgrave (trinikrono) wrote :

Maybe you need to press esc right after the BIOS loads when it says loading grub, if you are using grub legacy and not the newer grub 2.

Revision history for this message
kurt belgrave (trinikrono) wrote :
Revision history for this message
peterzay (peterzay) wrote :

I am using Grub2: grub-install (GNU GRUB 1.98-1ubuntu9).

Revision history for this message
peterzay (peterzay) wrote :

Just to recap. My Lenovo SL500 laptop runs Ubuntu 10.04 while my HP desktop test machine runs Ubuntu 10.10. Both experience the intermittent boot failure. Both use Grub2. Neither machine responds to Shift during the boot process. I am waiting for instructions on how to proceed. Thanks Kurt.

peterzay (peterzay)
Changed in ubuntu:
status: Incomplete → In Progress
Revision history for this message
peterzay (peterzay) wrote :

Please refer to Bug #425979 - Holding shift fails to display grub2 menu.

Changed in ubuntu:
status: In Progress → New
Revision history for this message
kurt belgrave (trinikrono) wrote :

Sorry for the delayed answer perter, can you press escape when you see the little ubuntu logo before the dots and see if you see the menu there.If that does not work try a Maverick cd and see if it gives any errors on the boot.
Cheers

Revision history for this message
peterzay (peterzay) wrote :

ESC does not give any menu for Grub2. A live CD boot can't nail down an intermittent problem.

Revision history for this message
kurt belgrave (trinikrono) wrote :

From the looks of things this bug is quiet common, what you can do is too
attach to this bug report the logfile /var/log/dmesg.0
And any other log that will help to pin down this issue.
Cheers

Changed in ubuntu:
status: New → Incomplete
Changed in ubuntu:
importance: Undecided → Low
Revision history for this message
kurt belgrave (trinikrono) wrote :
Revision history for this message
peterzay (peterzay) wrote :

You will find attached the log file /var/log/dmesg.0 as requested.

Revision history for this message
peterzay (peterzay) wrote :

I have just installed Boot Chart. The next time I get a boot failure, I will upload the audit graphic.

Revision history for this message
peterzay (peterzay) wrote :

Kurt, I sent the graphics you requested to your personal email to avoid publishing my computer name. Let me know if you also require the supplemental info zip files.

Changed in ubuntu:
status: Incomplete → In Progress
Revision history for this message
peterzay (peterzay) wrote :

Kurt, The boot failure happened again. I sent the graphics you requested to your personal email to avoid publishing my computer name. Let me know if you also require the supplemental info zip files.

Revision history for this message
kurt belgrave (trinikrono) wrote :

Peterzay
can you post the dmesg again
CHeers

Revision history for this message
kurt belgrave (trinikrono) wrote :

Peter
 You can go to /etc/default/grub and edit the line to remove "quiet splash"
after editing that file, use "sudo update-grub" in a terminal to make it take effect
Please post results

Changed in ubuntu:
status: In Progress → Incomplete
Revision history for this message
peterzay (peterzay) wrote :

You will find attached the log file /var/log/dmesg.0 as requested.

Revision history for this message
peterzay (peterzay) wrote :

Here is the post of the update-grub:

xxx:~$ sudo gedit /etc/default/grub
[sudo] password for xxx:
xxx:~$ sudo update-grub
Generating grub.cfg ...
Found linux image: /boot/vmlinuz-2.6.32-28-generic
Found initrd image: /boot/initrd.img-2.6.32-28-generic
Found linux image: /boot/vmlinuz-2.6.32-27-generic
Found initrd image: /boot/initrd.img-2.6.32-27-generic
Found linux image: /boot/vmlinuz-2.6.32-26-generic
Found initrd image: /boot/initrd.img-2.6.32-26-generic
Found linux image: /boot/vmlinuz-2.6.32-25-generic
Found initrd image: /boot/initrd.img-2.6.32-25-generic
Found linux image: /boot/vmlinuz-2.6.32-24-generic
Found initrd image: /boot/initrd.img-2.6.32-24-generic
Found linux image: /boot/vmlinuz-2.6.32-23-generic
Found initrd image: /boot/initrd.img-2.6.32-23-generic
Found linux image: /boot/vmlinuz-2.6.32-22-generic
Found initrd image: /boot/initrd.img-2.6.32-22-generic
Found linux image: /boot/vmlinuz-2.6.32-21-generic
Found initrd image: /boot/initrd.img-2.6.32-21-generic
Found memtest86+ image: /boot/memtest86+.bin
done
xxx:~$

Changed in ubuntu:
status: Incomplete → In Progress
Revision history for this message
kurt belgrave (trinikrono) wrote :

Peter
Is working for you now after you ran the commands?
Also i was speaking to the bugsquad about your error and it seems that it is fairly common, and a easy fix since all you have to do is restart

quote:

charlie-tca This is a not uncommon issue with both upstart and plymouth 19:16
charlie-tca sometimes with parallel processing during boot, things collide, causing the boot to fail 19:17
trinikrono charlie-tca: is that documented on a wiki anywhere? 19:18
charlie-tca I have no idea. It is a bug, but low priority 19:20
charlie-tca Easy workaround; reboot

Lastly
Please dont set the bug status to in progress ( that is for when a developer is working on a bug).
And do you have a package named flush installed you can try removing that and see if you boot fine since it is the last executeable i see running in bootchart.

Cheers

Revision history for this message
kurt belgrave (trinikrono) wrote :

Changing status
You can also remove the bootchart if you wish

Changed in ubuntu:
status: In Progress → Incomplete
Revision history for this message
peterzay (peterzay) wrote :

Kurt,

Yes, the boot process is now much more verbose. If I get the problem again, I am not sure to be able to capture the error message(s) manually. Can I dump this screen output into a file somehow?

For bug status, should I reset from incomplete to new when I reply?

No package called flush is installed on my system. The only search hit for flush on my file system is flush.pl and is attached.

I will remove BootChart.

Revision history for this message
peterzay (peterzay) wrote :

Both the failed boot and the good boot end with a flush-8:16 process. I don't think anything can be concluded from flush in this boot failure.

Revision history for this message
peterzay (peterzay) wrote :

Here is the good boot bootchart graphic.

Revision history for this message
peterzay (peterzay) wrote :

The priority of this bug is low??? An OS that can't boot fails in its most basic function. Sure a manual reboot is easy. What about remote environments?

Changed in ubuntu:
status: Incomplete → New
Revision history for this message
peterzay (peterzay) wrote :

It has been a month since I commented out the quiet splash from /etc/default/grub without any boot hangs. I am now enabling quiet splash.

Revision history for this message
peterzay (peterzay) wrote :

I just got a boot hang and will comment out quiet splash. The test period will be lengthened to 3 months.

Revision history for this message
kurt belgrave (trinikrono) wrote :

Peter can you look at this wikipage, i would like to make sure the bug is related to plymouth before i reassign.
https://wiki.ubuntu.com/Plymouth#Debugging

And please post back results hopefully you will have debug mode on when it fails to boot so you can see the root cause, take a picture if need be.
Cheers

Revision history for this message
kurt belgrave (trinikrono) wrote :

Setting to incomplete until you respond

Changed in ubuntu:
status: New → Incomplete
Revision history for this message
kurt belgrave (trinikrono) wrote :

Also please look at https://bugs.launchpad.net/ubuntu/+source/plymouth/+bug/609845
and tell me if this looks similar to what you are experiencing

Revision history for this message
peterzay (peterzay) wrote :

Wiki article https://wiki.ubuntu.com/Plymouth#Debugging requires entering the Grub2 boot menu. As indicated previously (Bug #425979), a Grub2 bug makes this impossible.

This boot failure bug appears to be similar to Bug #609845.

A curious feature of my bug is that when you comment out "quiet splash" in /etc/default/grub and then "sudo update-grub", the boot failure appears to be resolved.

I will report back after 3 months of verbose booting as indicated above, unless I upgrade to 11.04 Natty Narwhal.

In the interim, I would suggest that the 9 victims of Bug #609845 try verbose booting. If the problem goes away for all 9, then we can use that to refine our debugging efforts.

Changed in ubuntu:
status: Incomplete → New
Revision history for this message
peterzay (peterzay) wrote :

Kurt, I would suggest you wait for feedback from the 9 victims of Bug #609845 before reassigning my bug as a duplicate.

Revision history for this message
kurt belgrave (trinikrono) wrote :

Assigning to plymouth, maybe someone can better advise here
Cheers

affects: ubuntu → plymouth (Ubuntu)
Revision history for this message
kurt belgrave (trinikrono) wrote :

Please execute the following command, as it will automatically gather debugging information, in a terminal:
apport-collect 701060
let me know when you do

Changed in plymouth (Ubuntu):
status: New → Incomplete
Revision history for this message
peterzay (peterzay) wrote : BootDmesg.txt

apport information

tags: added: apport-collected
description: updated
Revision history for this message
peterzay (peterzay) wrote : CurrentDmesg.txt

apport information

Revision history for this message
peterzay (peterzay) wrote : Dependencies.txt

apport information

Revision history for this message
peterzay (peterzay) wrote : Lspci.txt

apport information

Revision history for this message
peterzay (peterzay) wrote : Lsusb.txt

apport information

Revision history for this message
peterzay (peterzay) wrote : ProcCpuinfo.txt

apport information

Revision history for this message
peterzay (peterzay) wrote : ProcInterrupts.txt

apport information

Revision history for this message
peterzay (peterzay) wrote : ProcModules.txt

apport information

Revision history for this message
peterzay (peterzay) wrote : UdevDb.txt

apport information

Revision history for this message
peterzay (peterzay) wrote : UdevLog.txt

apport information

Changed in plymouth (Ubuntu):
status: Incomplete → New
Revision history for this message
peterzay (peterzay) wrote :

Kurt,

The apport collect process has completed. Let me know if you need anything else.

Revision history for this message
Steve Langasek (vorlon) wrote :

When this bug is reproduced and you do get "Ubuntu with the luminous dots", do the dots continue to animate or are they fixed? Does pressing Esc at this point have any effect?

Revision history for this message
peterzay (peterzay) wrote :

Steve,

It has been quite some time since I experienced this bug frequently. I can't remember animation details and I have never tried ESC there.

This bug affects my production machine. I am very pleased, so far, by the workaround to comment out "quiet splash" in /etc/default/grub and then "sudo update-grub". The workaround has never failed.

Do you have a test machine where you can reproduce this problem?

If not, I can reset to "quiet splash" and wait and hope I am in front of my machine at the critical point.

Please let me know.

Revision history for this message
Steve Langasek (vorlon) wrote : Re: [Bug 701060] Re: Boot failure

On Wed, May 25, 2011 at 01:25:36AM -0000, peterzay wrote:
> It has been quite some time since I experienced this bug frequently. I
> can't remember animation details and I have never tried ESC there.

To advance on this bug, we will need to know these details.

> This bug affects my production machine. I am very pleased, so far, by
> the workaround to comment out "quiet splash" in /etc/default/grub and
> then "sudo update-grub". The workaround has never failed.

I'm happy that you have a workaround that works for you, but I'd rather we
fix the underlying bug. :)

> Do you have a test machine where you can reproduce this problem?

No, this issue is not reproducible for the vast majority of people. It may
be specific to your hardware, or specific to some software you have
configured at boot time. This is what we need to determine.

I see that your machine is listed as a Lenovo ThinkPad SL500, which is a
relatively uncommon model of ThinkPad. The more you can help us trace
through this on your hardware, the better we'll be able to pin down the bug,
I expect.

--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/
<email address hidden> <email address hidden>

Revision history for this message
peterzay (peterzay) wrote :

I will enable "quiet splash" in /etc/default/grub and then "sudo update-grub" and keep you posted.

Revision history for this message
peterzay (peterzay) wrote :

On a normal boot, ESC does nothing when the Ubuntu screen displays white dots that animate to red.

Revision history for this message
Steve Langasek (vorlon) wrote :

On Thu, May 26, 2011 at 01:30:27PM -0000, peterzay wrote:
> On a normal boot, ESC does nothing when the Ubuntu screen displays white
> dots that animate to red.

Uhm, that's very strange - first, because pressing 'esc' should break out
from the splash screen to the text console and show you the various boot
messages; second, because the dots are never supposed to be red but orange
:) I don't know if the wrong color is related to the other problems you're
having, but we definitely need to figure out why the esc key doesn't work in
order to debug this further.

Do Alt+Left, Alt+Right, or Alt+Fn keys work at this point?

Could you try doing the following after a successful boot?:

 - if logged in, log out
 - switch to VT1 with Ctrl+Alt+F1
 - log in as root (or log in as admin user and run sudo -s, if you don't have
   a root password)
 - run 'service gdm stop' to shut down X
 - run 'service plymouth start' to start the plymouth daemon
 - run 'plymouth splash; sleep 30; plymouth quit' as a single commandline

This should bring up the plymouth splash screen in a controlled environment
for 30 seconds at a time, letting you try to interact with it via the
keyboard and then dropping you back to the console after the timeout. Does
Esc have any effect if you do it this way?

--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/
<email address hidden> <email address hidden>

Revision history for this message
peterzay (peterzay) wrote :

Before trying your suggestions, have you reviewed Bug #425979 which documents problems with both SHIFT and ESC for Grub2 on certain machines?

As for orange vs red animated dots, that's probably screen colour balance and subjective interpretation.

What do you mean by: Do Alt+Left, Alt+Right, or Alt+Fn keys work at this point? Do you mean LeftAlt+Fn or RightAlt+Fn (n=1,2,3,...,12)? Please clarify.

Revision history for this message
peterzay (peterzay) wrote :

Are these procedures safe on a production machine that just has to work at all times?

 - if logged in, log out
 - switch to VT1 with Ctrl+Alt+F1
 - log in as root (or log in as admin user and run sudo -s, if you don't have a root password)
 - run 'service gdm stop' to shut down X
 - run 'service plymouth start' to start the plymouth daemon
 - run 'plymouth splash; sleep 30; plymouth quit' as a single commandline

Revision history for this message
Steve Langasek (vorlon) wrote :

On Sat, May 28, 2011 at 04:07:02PM -0000, peterzay wrote:
> Before trying your suggestions, have you reviewed Bug #425979 which
> documents problems with both SHIFT and ESC for Grub2 on certain
> machines?

I'm familiar with that issue; it's specific to key handling during early
boot (bootloader, before the kernel has started) and should not affect
keyhandling under plymouth.

> As for orange vs red animated dots, that's probably screen colour
> balance and subjective interpretation.

> What do you mean by: Do Alt+Left, Alt+Right, or Alt+Fn keys work at this
> point? Do you mean LeftAlt+Fn or RightAlt+Fn (n=1,2,3,...,12)? Please
> clarify.

These keys should allow you to switch VTs, away from plymouth to either a
blank screen or (possibly) a login prompt. This is LeftAlt + the Function
keys (primarily F1-F7), yes.

On Sat, May 28, 2011 at 04:14:15PM -0000, peterzay wrote:
> Are these procedures safe on a production machine that just has to work
> at all times?

> - if logged in, log out
> - switch to VT1 with Ctrl+Alt+F1
> - log in as root (or log in as admin user and run sudo -s, if you don't have a root password)
> - run 'service gdm stop' to shut down X
> - run 'service plymouth start' to start the plymouth daemon
> - run 'plymouth splash; sleep 30; plymouth quit' as a single commandline

There is always some risk when debugging a process that works with the
kernel framebuffer (as plymouth does) that you will lock up the display or
the whole system and have to reboot to restore access. And of course you
won't be able to keep any desktop applications running if you've logged out
and stopped gdm.

--
Steve Langasek Give me a lever long enough and a Free OS
Debian Developer to set it on, and I can move the world.
Ubuntu Developer http://www.debian.org/
<email address hidden> <email address hidden>

Revision history for this message
peterzay (peterzay) wrote :

During a normal boot, after the first animated white dot turns red, the following was observed:

- pressing SHIFT once: no effect
- pressing ESC once: takes you to a verbose screen as if quiet splash was commented out and then offers the usual multi user login screen
- LeftAlt+F1: gives a black screen with a small blinking underscore in the upper left corner, followed by the usual multi user login screen
- RightAlt+F1: same as LeftAlt+F1

Note that each of the above 4 tests were conducted with 1 boot, so 4 distinct boots were required to cycle through the tests.

I will wait for your feedback before testing the rest:

 - if logged in, log out
 - switch to VT1 with Ctrl+Alt+F1
 - log in as root (or log in as admin user and run sudo -s, if you don't have a root password)
 - run 'service gdm stop' to shut down X
 - run 'service plymouth start' to start the plymouth daemon
 - run 'plymouth splash; sleep 30; plymouth quit' as a single commandline

Revision history for this message
peterzay (peterzay) wrote :

I was testing my system for IPv6 as follows using admin account:
- set IPv4 from automatic to link-local only
- set IPv6 to automatic
- shutdown
- boot
- got check disk
- don't remember if I got animated red dots (sorry, was focussed on other things)
- got boot hang with black screen

Tried:
- ESC 1x and repeatedly: no effect
- SHIFT 1x and repeatedly: no effect
- CTRL+ALT+F1: no effect
- LeftAlt+F1: no effect
- RightAlt+F1: no effect

Then tried CTRL+ALT+DEL
- got a blinking screen with Ubuntu red dots superimposed on top of verbose splash messages
- that led to a successful reboot

Revision history for this message
peterzay (peterzay) wrote :

3 months have elapsed since I commented out the quiet splash from /etc/default/grub without any boot hangs. On this SL500 Lenovo laptop, the workaround resolves the boot hang problem.

Revision history for this message
dino99 (9d9) wrote :

That version has died long ago; no more supported

Changed in plymouth (Ubuntu):
status: New → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.