[jaunty][alpha6]X-Server restart randomly

[Lex Etyacc]

I run on a Dimension 9200, with and nvidia card.

Everything seems to run smoothly then the the screen just goes black for a second and I'm on the login screen again.

It sometimes seems linked to something I'm doing (hitting F5 to refresh a page on firefox, hiting Ctrl+S to save in gedit) but sometimes it just does it while I'm not doing anything.

It looks like the crashes come farther apart when I only use gnome-terminal and nothing else.

Let me know if I need to join more information from logs or stuff (I'm not sure what to put here, I probably shouldn't be running and alpha in the first place...)

That's the entry I found in syslog corresponding to the crashes I get
Mar 13 12:08:20 walhalla gdm[3836]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0

[Bremm (bremm)]

Same here, I've just found this now.

Mar 23 03:42:53 host gdm[4164]: WARNING: gdm_slave_xioerror_handler: Erro fatal no X - Reinicializando :0
Mar 19 21:14:38 host gdm[4770]: WARNING: gdm_slave_xioerror_handler: Erro fatal no X - Reinicializando :0
Mar 19 00:00:28 host gdm[4719]: WARNING: gdm_slave_xioerror_handler: Erro fatal no X - Reinicializando :0
Mar 17 03:44:19 host gdm[4660]: WARNING: gdm_slave_xioerror_handler: Erro fatal no X - Reinicializando :0
Mar 16 16:01:01 host gdm[4723]: WARNING: gdm_slave_xioerror_handler: Erro fatal no X - Reinicializando :0

It happens once after 1st login. If I reboot the system, it happens again once. The 2nd, 3rd, 4th, Nth... logins without a reboot don't zap.

Description: Ubuntu jaunty (development branch)
Release: 9.04

Linux host 2.6.28-3-rt #8-Ubuntu SMP PREEMPT RT Sun Mar 22 10:13:20 UTC 2009 x86_64 GNU/Linux

It was happening also with 2.6.28-2-rt. Each fatal error happened few minutes after each reboot. Side note: once I saw scroll lock led ON (and I didn't touch the key).

Following instructions at bug #346454, I'm attaching both X.Org logs.

[Vadim Peretokin (vperetokin)]

Same here, and I got two of these:

Mar 23 19:41:13 ubuntu x-session-manager[6313]: WARNING: Failed to send buffer

and then

Mar 23 19:41:18 ubuntu gdm[6087]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0

I attached the whole /var/log in #347642 (uploading atm), hopefully it has all the information necessary.

[Adal Alom Rodríguez (arpia49)]

It happens to me too.

gdm[2800]: WARNING: gdm_slave_xioerror_handler: ha ocurrido un error fatal de X - Reiniciando :0
gdm[18334]: WARNING: gdm_slave_xioerror_handler: ha ocurrido un error fatal de X - Reiniciando :0

BTW: "ha ocurrido un error fatal de X - Reiniciando :0" means "Fatal X error - Restarting :0" in spanish.

Dell Inspiron 1521 (AMD + ATI) with Jaunty.

If someone need more info... let me know.

[robe (r-evert)]

Same here, the only thing that shows up in the syslog is the X-restart.
Btw, I'm using Kubuntu 9.04.

[Vish (vish)]

same here , this is one log of several i have had!
gdm[16705]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0

the crashes are random and have occurred atleast 20-30 times since i'v installed jaunty a month ago...

using ATI mobility Radeon X1400... on an Acer 5672WLMi laptop...

this issue has been reported nearly 2 months ago, why is this still undecided? does any more info need to be provided to speed this bug resolution, because these random crashes are driving me crazy

[Thiago Teixeira (tvst)]

Same problem here. I'm going to mark this as confirmed, and hope someone in the right level notices it.

2.6.28-11-generic
ATI mobility radeon x1300 with the radeon driver
Lenovo Thinkpad T60 laptop

[Thiago Teixeira (tvst)]

oh, and in case it matters:

Atheros Communications Inc. AR5212 802.11abg NIC (rev 01)
with ath5k drivers

[Harima Shinji (playingshinji-deactivatedaccount)]

I am suffering from this X restart as well. Happens on random occasions.

shinji-desktop gdm[3249]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0

[Onaka (brollejunior)]

I'm suffering the same. On an ATI card though, an ATI radeon x1300 with unstable graphics drivers, but since the author experienced this and he has a card from NVIDIA I think it's not about the graphics driver.

[Jacob Winski (winski)]

Confirmed on an Intel video card. This is a MAJOR problem, because suspend/hibernate changes VT's, and a change in VT triggers this bug.

Easy way to trigger this bug: after logging into GNOME via GDM, press CTRL+ALT+2 to go to VT #2, then go back to the VT where you were logged into GNOME. It will have crashed, leaving you with a GDM login screen.

lspci of video card:
00:00.0 Host bridge: Intel Corporation Mobile PM965/GM965/GL960 Memory Controller Hub (rev 03)
00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 03)
00:02.1 Display controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 03)

Pruned /var/log/syslog output:
May 11 16:28:00 extensia kernel: [ 560.294382] [drm:i915_get_vblank_counter] *ERROR* trying to get vblank count for disabled pipe 0
May 11 16:28:00 extensia x-session-manager[4430]: WARNING: Detected that screensaver has left the bus
May 11 16:28:00 extensia gdm[4127]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0
May 11 16:28:00 extensia bonobo-activation-server (kuba-4984): could not associate with desktop session: Failed to connect to socket /tmp/dbus-hTpBtBgWnJ: Connection refused

[Jacob Winski (winski)]

*FIXED*

This is an issue with Xorg, specifically with the new option AccelMethod. Just remove or comment out the following line in your /etc/X11/xorg.conf file and the problem goes away:
Option "AccelMethod"

[Jacob Winski (winski)]

Notice: AccellMethod is NEVER in the xorg.conf file by default. It must be entered manually by the user, therefore no 'patch' can be released. Each user must fix it themselves.

[Adal Alom Rodríguez (arpia49)]

I can confirm it, there is no "Option "AccelMethod"" in my xorg.conf and it is still failing.

[GaryParr (gary-garyparr)]

No AccelMethod in my xorg.conf file and yet I get the random gdm_slave_xioerror_handler: Fatal X error.

Running 32 bit Jaunty release on an IBM R60 with ATI X1400 and open source driver. Using Emerald and Compiz. Have experienced this when starting VirtualBox guests, when opening a NetBeans project, and when just surfing the web on FireFox. There does not seem to be a common thread that ties them all together.

Did not have this issue on Hardy with proprietary ATI driver. Did I have other issues then? Yes, but nothing that killed my X session and caused me to lose data in running apps.

So, what's the "Fix" that was "Commited" on this bug? Did I miss something in the above posts?

[Vish (vish)]

@Kuba >

what do u mean no * fix can NOT * be released?
there is no line about AccelMethod in my xorg too..
how do u expect user to comment a line not present in the file?

[Thiago Teixeira (tvst)]

I don't think Kuba had the same problem as stated in this bug report. For one thing, I can switch to VTs just fine with Ctrl+Alt+F1, etc. Also, I do not have AccelMethod anywhere in my xorg.conf.

So I'm going to assume his fix doesn't apply to this report.

Kuba: Thanks for your help, but what makes you think the proposed fix is the solution? Do you have a link to an upstream bug report or mailing list discussion?

[Vish (vish)]

has this bug been fixed?
i have not had a crash for a lot of days.
previously i kept having crashes almost every day , sometimes more than twice.

has it been fixed for all ? or do others still have crashes?

[cviorel (viorel-ciucu)]

I can confirm this on my Thinkpad T61 with nVidia Quadro NVS 140M graphic card, running Ubuntu 9.04 x86 with Compiz and no 'AccellMethod' in xorg.conf. I experienced random crashes of the X server. I have gdm[3660]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0 in my /var/log/syslog.
I installed the latest driver from nvidia and my problems are gone!

[event.riga (event-riga)]

same issue here with Jaunty on Toshiba Satellite U400 and Intel G45 and Intel driver 2.7 and AccelMethod set to uxa but only when waking up from hibernate. Had no problem with Interpid and with 2.4 driver on Jaunty.

[mp (m-p)]

There is a lot of related reports - recent comments on this https://bugs.launchpad.net/bugs/355155 appears to be very similar. There are severe problems with xserver 1.6

But nothing is apparently done to fix it - not even a recognition or acknowledgement!?!?!?

Many people are switching away from Jaunty, because it is plainly unusable for a lot of people.

[csauer (caetanosauer)]

Has anybody experienced this bug with desktop effects disabled?

[event.riga (event-riga)]

No problems w/o compiz on my side.

[knb (knb)]

I have this problem, too. This week I lost critical data twice already.

My desktop effects are disabled via System/Preferences/Appearance/Visual Effects
However I also have this compizconfig applet installed, and some effects are indeed enabled thru *this applet* (does it matter?)

Linux 2.6.28-11-generic #42-Ubuntu SMP Fri Apr 17 01:58:03 UTC 2009 x86_64 GNU/Linux

02:00.0 VGA compatible controller: nVidia Corporation NV44 [GeForce 6200 TurboCache(TM)] (rev a1)

This never happened wih my old display, a 17 inch monitor. Recently I added a second 24 Inch monitor. Can it be tht the video card is just overwhelmed by providing signals to two displays? I'M just doing desktop work, though. Nothing graphics intensive.

I can't switch to a VT console with control+alt+f2. Screen stays blank.

[GaryParr (gary-garyparr)]

I recently spent some time "tinkering" with Compiz and Emerald and I haven't had the problem since. I have no idea what really changed since I ended up back where I started with both Compiz and Emerald enabled and the same set of Compiz options.

Don't you just love a consistent problem?

I would say that I was going to "tinker" again and see if I could try to zero in on some sort of cause for this issue, but to be honest... I think I'd much prefer having a stable system for a little while :-)

[[e-mail] (akoumjian)]

I'm having the same problem on my 2 machines running Jaunty:
1. HP Mini 2133
2. Whitebox with Asus A8N-Sli, NVidia Geforce 7900 GTX

Compiz is not enabled on the HP. It seems to be triggered when going into sleep.

[Bremm (bremm)]

This morning I gave a try on "nouveau" driver. It works as a charm (till now), and no more black screen after starting GDM for the first time after rebooting.

BTW, 190.09 in this PPA had problems with DKMS and 185.* stills with the same xorg zap bug.

I didn't test OpenGL games here yet, but gl and gl2 drivers @ mplayer sucked as usual. Before I was on vdpau and now back to xv driver.

Video card is 8500GT and kernel 2.6.28-11-generic (amd64).

[sperlyjinx (jgbaum)]

I have also been experiencing this nasty bug, but seemingly only since a recent kernel upgrade to 2.6.28-13-server. I haven't yet lost any critical data, but I have lost a lot of time. This bug has made my system unusable in a production environment which is disappointing, as I've been using Ubuntu as my main machine for the last three years. Here are the syslog entries from around the time of the latest crash:

Jul 21 11:19:42 toro gnome-session[20575]: WARNING: Detected that screensaver has left the bus
Jul 21 11:19:42 toro bonobo-activation-server (jgbaum-796): could not associate with desktop session: Failed to connect to socket /tmp/dbus-uDfPjG9umq: Connection refused
Jul 21 11:19:42 toro gdm[4931]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0
Jul 21 11:19:44 toro acpid: client 2477[0:0] has disconnected
Jul 21 11:19:44 toro last message repeated 3 times
Jul 21 11:19:44 toro acpid: client connected from 833[0:0]
Jul 21 11:19:46 toro last message repeated 3 times
Jul 21 11:19:56 toro pulseaudio[953]: pid.c: Stale PID file, overwriting.
Jul 21 11:19:57 toro pulseaudio[953]: module-x11-xsmp.c: X11 session manager not running.
Jul 21 11:19:57 toro pulseaudio[953]: module.c: Failed to load module "module-x11-xsmp" (argument: ""): initialization failed.
Jul 21 11:20:01 toro /USR/SBIN/CRON[1207]: (root) CMD ([ -x /usr/sbin/update-motd ] && /usr/sbin/update-motd 2>/dev/null)
Jul 21 11:20:01 toro /USR/SBIN/CRON[1210]: (root) CMD (if [ -x /usr/bin/vnstat ] && [ `ls /var/lib/vnstat/ | wc -l` -ge 1 ]; then /usr/bin/vnstat -u; fi)
Jul 21 11:20:12 toro python: hp-systray[997]: error: Unable to lock /home/jgbaum/.hplip/hp-systray.lock. Is hp-systray already running

FYI, I am running jaunty with 2 Nvidia Quadro NVS 290 cards on 3 displays.

[sperlyjinx (jgbaum)]

Here is the gdm log from around the time of the crash.

[Mike Butash (michael-butash)]

I've been experiencing this about every couple of days, and each time it seems related to my using the scroll wheel in firefox to whisk through pages. My mouse is a logitech mx revolution with a broken scrollwheel that scrolls indefinitely (good bearings) when spun - I've not had issues with scrolling really fast prior in ibex before updating, but perhaps an evdev bug with the device or rate of that input changing? GDM, syslog, messages, xorg logs all show very little other than the gdm_slave_xioerror_handler: fatal x error everyone else has.

Can anyone else say if they were scrolling with their mouse or anything like it at the same time their desktop was crashing?

[John Storta Jr. (jstorta)]

I had X crash on me twice today. Each time I got the following in the logs.

First crash
x-session-manager[3183]: WARNING: Detected that screensaver has left the bus
gdm[2743]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0
acpid: client 2748[0:0] has disconnected
acpid: client 2748[0:0] has disconnected
acpid: client connected from 28410[0:0]
acpid: client connected from 28410[0:0]

Second crash
x-session-manager[28445]: WARNING: Detected that screensaver has left the bus
gdm[28375]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0
acpid: client 28410[0:0] has disconnected
acpid: client 28410[0:0] has disconnected
acpid: client connected from 29576[0:0]
acpid: client connected from 29576[0:0]
acpid: client connected from 2700[111:120]

Here is my hardware.
Asus M3A motherboard
AMD Athlon 64 X2 4600+ CPU
3GB physical memory
XFX 8800 GT Alpha Dog 512MB DDR3 Standard (PV-T88P-YDF4) Video card with restricted driver (180)

The first message mentions the screensaver, but in both cases I was in the middle of doing stuff when I suddenly got dropped to a login screen. The screen saveer was not on. And I have screen saver set to just blank screen. Apps I had open included Firefox, terminal, and jEdit.

I looked in /etc/X11/xorg.conf and it does not contain 'AccellMethod' anywhere in it.

I saw someone mention that turning off the compiz extra effects is a solution. Is it the consensus from folks on this thread that users should not run the Extra Effects? That seems more like a workaround rather than a solution.

If anyone has anything else that I might try to isolate this problem, please let me know what I can do.

Thanks,
John S.

[John Storta Jr. (jstorta)]

X just crashed again. I had a bunch of apps open.

I was posting a message on a message board (not this one) when it crashed.

From daemon.log
Aug 20 15:16:56 aragorn x-session-manager[3195]: WARNING: Detected that screensaver has left the bus
Aug 20 15:16:56 aragorn gdm[2757]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0
Aug 20 15:16:57 aragorn acpid: client 2761[0:0] has disconnected
Aug 20 15:16:57 aragorn acpid: client 2761[0:0] has disconnected
Aug 20 15:16:57 aragorn acpid: client connected from 12151[0:0]
Aug 20 15:16:59 aragorn acpid: client connected from 12151[0:0]

From messages
ug 20 15:16:57 aragorn bonobo-activation-server (john-12111): could not associate with desktop session: Failed to connect to socket /tmp/dbus-SUkOtBChGD: Connection refused
Aug 20 15:17:09 aragorn pulseaudio[12332]: pid.c: Stale PID file, overwriting.

From syslog
Aug 20 15:16:56 aragorn x-session-manager[3195]: WARNING: Detected that screensaver has left the bus
Aug 20 15:16:56 aragorn gdm[2757]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0
Aug 20 15:16:57 aragorn bonobo-activation-server (john-12111): could not associate with desktop session: Failed to connect to socket /tmp/dbus-SUkOtBChGD: Connection refused
Aug 20 15:16:57 aragorn acpid: client 2761[0:0] has disconnected
Aug 20 15:16:57 aragorn acpid: client 2761[0:0] has disconnected
Aug 20 15:16:57 aragorn acpid: client connected from 12151[0:0]
Aug 20 15:16:59 aragorn acpid: client connected from 12151[0:0]

Just wanted to add the latest details.

Thanks

[John Storta Jr. (jstorta)]

I found the solution to my problem.

I was having issues with my system crashing occassionally, X crashing occassionally, and Flash causing Firefox to hang.

After a great deal of investigation and significant assistance from the Ubuntu community, I found a BIOS setting called 'AMD Cool n Quiet'. It does something with regard to CPU power management.

I disabled this "feature" and I have not had a single problem since. I have done everything imaginable to stress my system and reproduce any of the issues. Nothing.

System has not crashed. X has not crashed. I have played hundreds of Flash videos in Firefox and I have not had a single hang.

Just wanted to update the thread with what solved the problem for me just in case it applies to someone else as well.

Thanks,
John S.

[hkais (r-2)]

Hello all,

Same issue here. I hava a Lenovo T61 with nvidia Quadro NVS 140M.
I have no AMD and therefore no AMD Cool n Quiet. It is pretty anoying to get the work kill randomly.

Is there an solution also for intel-bases systems like mine?

[Mike Butash (michael-butash)]

I think _every_ time my system crashes, it's been through an input event, namely when I'm scrolling through a web page. After a few times of it happening, I thought about it, and just about every time since, it's the same. Can anyone else correlate what they're doing to when it crashes? I suspect a problem with input events in Xorg, but my logs are as vague as anyone else's to point a finger...

Dating back to Edgy, I've had a perpetual memory leak in X or the kernel that I think is related to the NV proprietary drivers ( I see in 64bit or 32 systems), but this is the first time in a long time I've had X simply crash this violently and abruptly. I think it might be adding to the frequency of the crashes, about every 3 or so days here.

Thanks!

[cviorel (viorel-ciucu)]

The error reappeared even after nvidia driver update on my ThinkPad T61 with nVidia Quadro NVS 140M graphic card, running Ubuntu 9.04.

[hkais (r-2)]

I think this is a duplicate of #342324

[hkais (r-2)]

Bug #346454 and Bug #412113