X crashes after pressing enter or 2 in 10.04 final release

[petrj]

Binary package hint: plymouth

This bug looks like the bug #532047 in beta-1. I have installed Lucid Lynx final release (system is fully updated right now) and this bug persists :-( Whenever I boot and press <Enter> or <2> X crashes and restart, It's annoying !

ProblemType: Bug
DistroRelease: Ubuntu 10.04
Package: plymouth 0.8.2-2ubuntu2
ProcVersionSignature: Ubuntu 2.6.32-21.32-generic 2.6.32.11+drm33.2
Uname: Linux 2.6.32-21-generic x86_64
NonfreeKernelModules: h826d nvidia
Architecture: amd64
Date: Wed May 12 15:44:08 2010
DefaultPlymouth: /lib/plymouth/themes/kubuntu-logo/kubuntu-logo.plymouth
InstallationMedia: Kubuntu 10.04 "Lucid Lynx" - Release Candidate amd64 (20100419.1)
MachineType: ATComputers PC ALZA
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-21-generic root=UUID=53db1a11-a4a4-4f63-9ddb-4036ee7bc0e2 ro
ProcEnviron:
 LANGUAGE=
 LANG=cs_CZ.UTF-8
 SHELL=/bin/bash
ProcFB:
 0 VGA16 VGA
 1 VESA VGA
SourcePackage: plymouth
TextPlymouth: /lib/plymouth/themes/ubuntu-text/ubuntu-text.plymouth
dmi.bios.date: 07/17/2008
dmi.bios.vendor: Award Software International, Inc.
dmi.bios.version: F5
dmi.board.name: GA-MA78GM-S2H
dmi.board.vendor: Gigabyte Technology Co., Ltd.
dmi.chassis.asset.tag: DL006497
dmi.chassis.type: 3
dmi.chassis.vendor: ATComputers
dmi.chassis.version: Alza GameBox 5000+ bez OS
dmi.modalias: dmi:bvnAwardSoftwareInternational,Inc.:bvrF5:bd07/17/2008:svnATComputers:pnPCALZA:pvrAlzaGameBox5000+bezOS:rvnGigabyteTechnologyCo.,Ltd.:rnGA-MA78GM-S2H:rvr:cvnATComputers:ct3:cvrAlzaGameBox5000+bezOS:
dmi.product.name: PC ALZA
dmi.product.version: Alza GameBox 5000+ bez OS
dmi.sys.vendor: ATComputers

[Steve Langasek (vorlon)]

 status incomplete

On Wed, May 12, 2010 at 01:51:16PM -0000, petrj wrote:
> I have installed Lucid Lynx final release (system if fully updated right
> now) and this bug persists :-( Whenever I boot and press <Enter> or <2>
> X crashed and restart, It's annoying !

Please reboot and set the option 'plymouth:debug' on the kernel command line.

After booting, please switch to console with Ctrl+Alt+F1, log in, and report
the output of 'ps awx | grep X' as well as the output of 'status plymouth';
then switch back to X with Alt+F7 and confirm that the crash happens on the
same boot cycle.

Finally, attach the /var/log/plymouth-debug.log file that's created on the
system.

--
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>

[petrj (petrjanousek-net)]

Thank you for you time, Steve,

I've tried to reboot almost 10 times and I've found that when I boot with 'ro plymouth:debug', very often X doesn't start, (attachment plymouth-debug.log.failedstart'). Sometimes X starts, with no crash after <enter> or <2> (attachment plymouth-debug.log.ok)
When I boot wih 'ro splash plymouth:debug', it;s possible to simulate crash after <enter> or <2>. (attachment 'plymouth-debug.log.crash).
'ps awx | grep X' output:

 2172 tty7 Ss+ 0:22 /usr/bin/X -nr -nolisten tcp :0 vt7 -auth /var/run/xauth/A:0-iySN9a
 2298 ? S 0:00 kdeinit4: kio_trash [kdeinit] trash local:/tmp/ksocket-kacenka/klauncherhX2277.slave-socket local:/tmp/ksocket-kacenka/plasma-desktopbJ2294.slave-socket
 2299 ? S 0:00 kdeinit4: kio_file [kdeinit] file local:/tmp/ksocket-kacenka/klauncherhX2277.slave-socket local:/tmp/ksocket-kacenka/plasma-desktopVf2294.slave-socket
 2300 ? S 0:00 kdeinit4: kio_file [kdeinit] file local:/tmp/ksocket-kacenka/klauncherhX2277.slave-socket local:/tmp/ksocket-kacenka/plasma-desktopFe2294.slave-socket
 2308 ? S 0:00 kdeinit4: kio_http [kdeinit] http local:/tmp/ksocket-kacenka/klauncherhX2277.slave-socket local:/tmp/ksocket-kacenka/plasma-desktopcG2294.slave-socket
 2309 ? S 0:00 kdeinit4: kio_http [kdeinit] http local:/tmp/ksocket-kacenka/klauncherhX2277.slave-socket local:/tmp/ksocket-kacenka/plasma-desktopNk2294.slave-socket
 2784 pts/2 S+ 0:00 grep --color=auto X

'status plymouth' output:

plymouth stop/waiting

[petrj (petrjanousek-net)]

plymouth-debug.log.ok

[petrj (petrjanousek-net)]

plymouth-debug.log.crash

[Steve Langasek (vorlon)]

Thanks. Can you try installing the gdm package, and checking whether the problem is reproducible with gdm instead of kdm? This may be a bug in the integration between kdm and plymouth.

[petrj (petrjanousek-net)]

I completly removed kdm and replaced with gdm. Still the same crash. When I pressed <enter> at the gdm login screen, X crashed and restarted. (I set up autologin in kdm, so the crash appears not in login screen but in the kde desktop).

Plymouth-debug.log created while starting with gdm is in attachment.

[petrj (petrjanousek-net)]

I would appreciate some fix or workaround, please ! Am I alone who is affected with this bug ?

[Steve Langasek (vorlon)]

On Thu, May 27, 2010 at 10:05:16AM -0000, petrj wrote:
> I would appreciate some fix or workaround, please ! Am I alone who is
> affected with this bug ?

It really appears that you are. There were bugs about this much earlier in
the development cycle that affected a huge number of our developers, but
those were all resolved; I don't know why those problems would affect you,
we thought we solved them at their root.

--
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>

[petrj (petrjanousek-net)]

Is it possible to use another version of plymouth package (older version, backport or something similar) ? Or is there any way to change boot process behaviour to avoid this bug ?

[euthymos (euthymos)]

Terrible bug please FIX!!!!

[petrj (petrjanousek-net)]

Please fix it !

[petrj (petrjanousek-net)]

I installed Lucid Lynx to another partition on my HD and I've found that X server is not restarting after ENTER or 2.
Then I installed Maverick Meerkat alpha 2 release and the same behaviour - no restart after ENTER or 2.

(I tried also to replace plymouth package with the version from Maverick Meerkat - and X is crashing after enter)

There must be something wrong in my plymouth//X/KDE configuration ! Could someone help me, please ? I don't wanna reinstall it !

[petrj (petrjanousek-net)]

I found the reason of this bug. I wrote PHP cli service that is starting during boot. ( by update.rc xxxx.php defaults ). When I remove this service, booting is OK. I think that the problem is in PHP "echo" function, but I'm not sure.

<irony>Thank you for your help, community </irony>

[Steve Langasek (vorlon)]

Can you provide a copy of this service so that we can try to reproduce the error?

> <irony>Thank you for your help, community </irony>

... Did you expect us to be able to debug this issue without knowing anything about this init script no one else has that triggers it?

[petrj (petrjanousek-net)]

I attached example of PHP service that evokes the restart error.

First file "exemplary_service" is startup script from /etc/init.d/.
Second file "exemplary_service.php" is PHP cli script from /var/www/

Service was installed by "update-rc.d exemplary_service".

I installed exactly the same service to my work notebook, but no restart bug occured (upgraded to Lucid Lynx from Karmic Koala last week) My notebook is HP Elitebook 8530p with ATI Graphics Card. The PC where the bug occures is settled with NVIDIA 9600 GT. Is it possible that the bug relates with NVIDIA graphic card ?

Sorry about the </irony> tag.

[Launchpad Janitor (janitor)]

[Expired for plymouth (Ubuntu) because there has been no activity for 60 days.]

[Colin Watson (cjwatson)]

@petrj: I'm pretty sure that the interaction with your PHP service is a coincidence. I believe this is a race condition, so changing random things to arrange for more or fewer processes to be running could easily perturb the behaviour. This behaviour of X happens when the terminal X is running on is set into Canonical Mode under X's feet: https://bugs.launchpad.net/ubuntu/+source/plymouth/+bug/540256/comments/5 is a detailed description of a sequence of events from a previous occurrence of this bug, although that bug has been fixed so the details will not be quite the same here.

I've been investigating an strace given to me by Naty Bidart, who was experiencing a similar bug. I can only make some guesses at what's happening as this strace only covers what plymouthd is doing:

 * plymouth sets terminal into Raw Mode on startup
 * plymouth appears to be configured to use the details theme (i.e. no splash), which is not the same as @petrj's debug log - in fact, in @petrj's case, things work *only* if using the details theme
 * gdm calls plymouth deactivate to prepare for transition to X, which sets terminal back into Canonical Mode
 * gdm calls plymouth --has-active-vt, which says yes, so it then calls gdm_server_start_on_active_vt
 * about 0.4 seconds later, gdm calls plymouth quit --retain-splash, which cleans up and exits
 * at no point does plymouthd change the active VT, implying that http://cgit.freedesktop.org/plymouth/commit/?id=8c7b2de6913870167bee60f73423841721f1a2cc (which initially looked like a promising fix) is not relevant

Unfortunately this doesn't leave me a whole lot wiser than before, but I thought it was better to post this than not. Perhaps it will trip a switch in somebody's mind, or maybe I will come back to this later and figure it out ...

[Eric Appleman (erappleman)]

still getting this in oneiric live sessions

[Eliah Kagan (degeneracypressure)]

Indeed, in all recent Oneiric daily-lives, this occurs when "Try Ubuntu without installing" is selected from the text-based boot menu, but not when that menu is bypassed and "Try Ubuntu" is selected in the graphical ubiquity-dm dialog. Furthermore, after a complete installation is carried out with Ubiquity, the graphical live session stops exhibiting this behavior. There might be additional intricacies that I have not identified. Please see duplicate bug 798509 for details. I have tested this on two machines of different hardware, and Eric Appleman has posted here confirming for Oneiric as well. So far no one has reported following the above instructions on Oneiric and failing to reproduce this bug--it would probably be good of more people could test, to ensure this is really happening for everyone in Oneiric (as it appeared that it wasn't, in Lucid).

I am not sure if this is really a bug in plymouth, at least as it occurs in Oneiric. But I don't consider myself qualified to tell. In duplicate bug 798509, I posted some logs that I believe indicate this may be a problem with key mapping in X.

I've changed this bug's title and tagged it to make clear that this affects Oneiric. If no one objects soon, I'll edit the description to reflect that as well.

[Steve Langasek (vorlon)]

I have not reproduced this issue in 10.04 final release, but I know for certain that the crashes in oneiric dailies is a different bug. The bug there is that lightdm, the new DM being introduced in oneiric as default, does not integrate properly with plymouth. A new bug should be filed against lightdm about this.

[Eric Appleman (erappleman)]

I have a persistent stick, I'll see if uninstalling lightdm and using gdm instead is possible.

I'll try to initiate a crash on the ubiquity approach.

[Eliah Kagan (degeneracypressure)]

The information I had posted here was based on Timo Aaltonen (https://launchpad.net/~tjaalton) having marked bug 798509 as a duplicate of this bug. Since bug 798509 is about the Oneiric-specific issue where X terminates when 2 (on the alphanumeric portion of the keyboard) or enter (either enter key) is pressed, it seems to me that no *new* bug has to be reported; that bug could simply be unmarked as a duplicate of this bug.

Based on Steve Langasek's post (https://bugs.launchpad.net/ubuntu/+source/plymouth/+bug/579313/comments/22), I have done just that--bug 798509 is no longer a duplicate of this bug (and I have added lightdm in Ubuntu as having bug 798509). I have posted there to explain this, and I have also subscribed Timo Aaltonen to this bug and both developers (Aaltonen and Langasek) to bug 798509, so that everybody knows what's going on, and nobody feels like they are being ignored. I apologize if this action (subscribing both of you) is inappropriate.

@Eric Appleman
I do not believe that the problem in Oneiric is actually a crash, per se--as explained in bug 798509, inspection of the dmesg on my machines doesn't reveal one, nor does inspection of the contents of /var/crash (both inspections conducted after X terminates). But by all means, if you're able to produce a crash related to this, please post about it. Or perhaps you just mean that you're going to try to get this problem to occur on an Oneiric system after selecting "Try Ubuntu" in the graphical ubiquity-dm menu. In any case, please note that since Steve Langasek has indicated that the problem in Oneiric is a different bug from this one, you should probably subscribe to bug 798509, indicate it affects you, and post any relevant results there, unless that bug gets re-marked as a duplicate of this one.

[digitalgecko (stephen-n)]

Hello Customer

You can prefer any watch for her that matches her taste and your resources.
We offer goods only for those of you who prefer and like the goods of good quality.

**************************************************************
Received my Rolex Daytona yesterday…best replica I have ever seen.
Thankee
                     Domingo Grover
**************************************************************

Click here ---> http://lastluxuryorder.ru

[dino99 (9d9)]

That version has died long ago; no more supported