Computer freezing how to troubleshoot?

m79reed · 2012-02-09 15:37:48

Today i have a lot of freezes and crashes.
It's freezing for 2-3 seconds (there is a looping sound when i'm listening music) and after that computer restart automatically.
On that 2-3 seconds time i'm NOT able to use capslock or numlock keys. Leds on keyboard are off.

Linux localhost 3.2.5-1-ARCH #1 SMP PREEMPT Tue Feb 7 08:34:36 CET 2012 x86_64 AMD Phenom(tm) II X4 945 Processor AuthenticAMD GNU/Linux
Ati Radeon 5770 with radeon driver, not catalyst
Gnome3

kernel starts with "Extreme Debug": https://wiki.archlinux.org/index.php/Boot_Debugging with no any logs...

Last edited by m79reed (2012-02-09 15:40:18)

m79reed · 2012-02-13 12:50:55

I've change a video driver from ati to catalyst and there is no any freezes. It's second day...
I suppose the problem is DRI and that errors in Xorg.log:

[  1676.990] (II) RADEON(0): radeon_dri2_flip_event_handler:1067 fevent[0x2835aa0] width 1920 pitch 7680 (/4 1920)
[  1676.992] (II) RADEON(0): radeon_dri2_schedule_flip:670 fevent[0x16520a0]

subatomic · 2012-02-28 07:45:46

The strangest thing happened the other day. The system was running smoothly for hours so I took a chance and left it on overnight, and turned the display off. In the morning I turned the display back on, moved the mouse slightly and was amazed to see it still running as I left it. I was sure the problem was solved until I returned 10 minutes later with my morning coffee to a frozen system!

subatomic · 2012-03-14 19:30:12

So has everyone "solved" this? My system continued to freeze randomly and I have been monitoring the hardware. At the last freeze the CPU was at 2.8%, temps were at 114F/45C and 102F/39C, RAM was at 0.9 and swap was 0. No logs as usual. File indexing has been off the entire time.

It took me a while but I eventually went with the linux-lts kernel and it may be too early to tell but I may have found my solution here.

AND my Bluetooth adapter is recognized again!

Last edited by subatomic (2012-03-14 20:17:44)

Lone_Wolf · 2012-03-18 12:44:28

Another option to help with troubleshooting (or have to chance to do a clean reboot) :

enable Magic SysRq key.

archwiki on sysctl
wikipedia onMagic_SysRq_key

dabd · 2012-03-18 16:39:18

I am running VirtualBox when this issue happens. I read that kernel developers consider VirtualBox's driver as "tainted crap" http://linux.slashdot.org/story/11/10/1 … river-crap.
Could it be the cause?

Wilco · 2012-03-19 08:42:12

Same problem here, see my topic about hard lockups. I though my crashes were related to the buggy broadcom-wl driver (the only driver that works due to a bug in brcmsmac) but seeing these comments I begin to wonder if there is a kernel/glibc bug somewhere.

ToxinPowe · 2012-03-19 13:42:13

Same problem here, my config:

Linux myhost 3.2.11-1-ARCH
NVIDIA driver 295.20 (8600GT)
KDE 4.8.1

Grima · 2012-03-20 07:29:06

I got the same problem too... My computer freezes randomly (I'm not doing anything specific) and I really do mean freeze. Mouse can't be moved, everything stops updating (on the screen), CTRL+ALT+F1-7 doesn't work and so on.

Right now I woke up with the gnome screensaver active but the computer had freezed during the night. The time in the screensaver was around 1 am. 7 hours of doing nothing in other words... So annoying!

local/linux 3.2.11-1
local/nvidia 295.20-3

Wilco · 2012-03-20 10:48:24

Something is definitely buggy, the only common denominator is we all use kernel 3.2

subatomic · 2012-03-23 08:35:04

that's what I thought until I installed the lts kernel (3.0.25) and nvidia-lts, unfortunately it freezes exactly as before although (so far) just when X is running.
I've tried downgrading suspected pacakages and adding things to the kernel line in grub, but now I am suspecting Xorg or perhaps powersaving.

ToxinPowe · 2012-03-23 15:05:58

I have disabled desktop effects on KDE, and I have no problems anymore.
Maybe the problem for me was the Nvidia 295.x driver?

I hope this can be useful for someone.

jcg · 2012-09-03 02:41:16

Not to revive a dead horse (and then beat it), but this is happening regularly with my Lenovo W500 (Intel GM45 or Radeon M86GL) running the 3.5.3 kernel. Is anyone else experiencing the issue? Has anyone resolved it? I was very happy with Arch until this started happening, now I'm thinking I'll have to switch back to Debian... the horror.

Last edited by jcg (2012-09-03 19:11:16)

subatomic · 2012-09-03 06:41:04

What is "regularly"? When did it start, can you downgrade some packages? Stop unnecessary daemons and read your logs, eventually you will find it.
In my case it was lirc and it took me a few months to find it. It's like this.

ewaller · 2012-09-03 15:13:39

jcg wrote:

Not to revive a dead horse (and then beat it), but this is happening regularly ...

This is an old thread, and much has been said and suggested. Can you please clarify what is 'this' ?

Are you getting a kernel panic? (Keyboard LEDs flashing)
Can you change virtual terminals using Ctrl-Alt-Fn ?
Can you ssh into the box when it is frozen?
Can you ping the box when it is frozen?
Is the disk access light on solid?
Does pressing the Caps Lock key have any effect on the Caps Lock LED on the keyboard?

jcg · 2012-09-03 18:19:37

Thanks for the replies ewaller and subatomic. I'll try to address your questions in the following post, but first, apologies for the hysterical post, of course I don't intend to switch back to Debian, Arch ftw.

By ``regularly'' I meant that it happens approximately once every 24 hours of uptime, or so. Apparently random timing, usually (perhaps always?) triggered by a mouse click.

By ``this'' I meant the common denominator of what is described above. That is,

(1) No kernel panic blinky-lights.
(2) CAPS-LOCK and NUM-LOCK keys seem to work (the indicators light up).
(3) Wifi indicator light keeps blinking as if things are functioning normally.
(4) Disk activity light keeps blinking as normal.

Failure mode is instant black screen, except for the cursor, which is still rendered. However, the cursor is ``frozen'' in that it will not move with mouse inputs. CTRL-ALT-FN does nothing as far as I can see. I have attempted to login ``blind'' to e.g. CTRL-ALT-F1 tty1 in the vain hope that i might be able to reboot/shutdown the system or stop X that way, but no-go.

Interesting point: sleep and suspend keys still work as normal, but upon wake the system resumes its minimally-functional state.

The only way I know to ``recover'' is to hold down the power button.

I have attempted to ``fix'' this by running the LTS kernel. I am now 19m15s into my first LTS boot without any symptoms. Will report back if it happens on LTS, although I realize that isn't a fix at all.

I have tried looking through various /var/log/*.log files and found nothing surprising or clearly problematic, but I'll stay vigilant if it starts happening with the LTS kernel (or if I feel adventurous with the bleeding-edge kernel).

I don't have another machine to ssh into this one with, but I'm working on it. Just waiting for my HDMI cable to arrive for my Raspberry Pi. I'll try pinging and sshing if it happens with LTS kernel or, again, if I get adventurous.

ewaller · 2012-09-03 19:41:34

Well, the good news is that it is not a kernel panic.

The cursor is visible -- so it not a matter of no output or no backlight.
The cursor is probably generated in hardware, so X could be tango-uniform.

How are you starting X? You might consider turning off login managers, logging in at a console, and using startx (Which may require playing with ~/.xinitrc). This will allow stdout and stderr messages to be displayed on the console. Also, when things hang, you can try to Ctrl-Alt-F1 back to the console knowing that you are logged in already. Then, a Ctrl-C would kill the X session

Make sure you look though /var/log/Xorg.0.log for clues. Note that file is created anew each time X starts, so you wither need to look at it outside of X, or look at /var/log/Xorg.1.log, which is the previous session's log.

jcg · 2012-09-03 22:27:26

ewaller wrote:

Well, the good news is that it is not a kernel panic.
The cursor is visible -- so it not a matter of no output or no backlight.
The cursor is probably generated in hardware, so X could be tango-uniform.
How are you starting X? You might consider turning off login managers, logging in at a console, and using startx (Which may require playing with ~/.xinitrc). This will allow stdout and stderr messages to be displayed on the console. Also, when things hang, you can try to Ctrl-Alt-F1 back to the console knowing that you are logged in already. Then, a Ctrl-C would kill the X session
Make sure you look though /var/log/Xorg.0.log for clues. Note that file is created anew each time X starts, so you wither need to look at it outside of X, or look at /var/log/Xorg.1.log, which is the previous session's log.

Well, I can (un?)happily report that the LTS kernel did nothing to solve the problem. I have experienced it twice in the last ~4 hours.

I am starting X using the startx command. I don't have any login managers or desktop environment (using openbox exclusively).

As for switching to an already logged-in tty, interestingly enough, it does that automatically! I was running emacs in tty2, startx in tty1, X took a dump in `tty7' which sent me back to tty2. Unfortunately, the cursor was rendered over tty2 and I had neither control over anything in tty2 nor the ability to switch to another tty...

Also, my previous statement about CAPS-LOCK and NUM-LOCK working was erroneous. They don't.

Here is a listing of the output of lsmod and the last few entries of Xorg.1.log:

[user@(none) ~]$ lsmod
Module                  Size  Used by
snd_hda_codec_conexant    46762  1 
arc4                    1410  2 
iwlagn                227948  0 
snd_hda_intel          22448  2 
snd_hda_codec          79127  2 snd_hda_intel,snd_hda_codec_conexant
i915                  592729  2 
thinkpad_acpi          62736  0 
snd_hwdep               6364  1 snd_hda_codec
uvcvideo               64494  0 
pcmcia                 35913  0 
mac80211              215428  1 iwlagn
snd_pcm                74656  2 snd_hda_codec,snd_hda_intel
r852                   10796  0 
nvram                   5970  1 thinkpad_acpi
sm_common               7482  1 r852
i2c_algo_bit            5295  1 i915
snd_page_alloc          7217  2 snd_pcm,snd_hda_intel
acpi_cpufreq            5813  1 
videodev               78306  1 uvcvideo
nand                   45422  2 sm_common,r852
snd_timer              19126  1 snd_pcm
btusb                  12100  0 
snd                    57456  11 snd_timer,snd_pcm,snd_hwdep,thinkpad_acpi,snd_hda_codec,snd_hda_intel,snd_hda_codec_conexant
drm_kms_helper         26101  1 i915
mperf                   1275  1 acpi_cpufreq
cfg80211              162447  2 mac80211,iwlagn
sdhci_pci               8402  0 
nand_ecc                3572  1 nand
v4l2_compat_ioctl32     7460  1 videodev
firewire_ohci          29439  0 
sdhci                  22066  1 sdhci_pci
cpufreq_userspace       2144  0 
bluetooth             138302  1 btusb
soundcore               5634  1 snd
drm                   183594  3 drm_kms_helper,i915
nand_ids                5649  1 nand
tpm_tis                 8074  0 
media                  10245  2 videodev,uvcvideo
mmc_core               73469  1 sdhci
r592                   12011  0 
mtd                    25703  2 nand,sm_common
serio_raw               4326  0 
firewire_core          51274  1 firewire_ohci
mei                    31521  0 
tpm                    11557  1 tpm_tis
intel_agp              10904  1 i915
rfkill                 15668  4 bluetooth,cfg80211,thinkpad_acpi
psmouse                54902  0 
intel_gtt              13975  3 intel_agp,i915
tpm_bios                5121  1 tpm
iTCO_wdt               12493  0 
processor              24355  3 acpi_cpufreq
memstick                6984  1 r592
i2c_i801                8116  0 
wmi                     8475  0 
yenta_socket           21117  0 
evdev                   9338  7 
pcmcia_rsrc             8857  1 yenta_socket
iTCO_vendor_support     1929  1 iTCO_wdt
crc_itu_t               1363  1 firewire_core
battery                 6349  0 
ac                      2408  0 
button                  4502  1 i915
thermal                 8023  0 
video                  11307  1 i915
i2c_core               20433  6 i2c_i801,drm,drm_kms_helper,videodev,i2c_algo_bit,i915
pcmcia_core            12189  3 pcmcia_rsrc,yenta_socket,pcmcia
microcode              12431  0 
ext4                  371003  3 
crc16                   1359  2 ext4,bluetooth
jbd2                   67031  1 ext4
mbcache                 5817  1 ext4
sd_mod                 27931  5 
sr_mod                 14727  0 
cdrom                  35616  1 sr_mod
pata_acpi               3408  0 
ata_generic             3359  0 
uhci_hcd               23500  0 
ata_piix               22063  4 
libata                174174  3 ata_piix,ata_generic,pata_acpi
scsi_mod              132145  3 libata,sr_mod,sd_mod
ehci_hcd               40327  0 
usbcore               144398  5 ehci_hcd,uhci_hcd,btusb,uvcvideo

[ 25085.617] (**) Option "xkb_layout" "us"
[ 25123.441] (II) evdev: Power Button: Close
[ 25123.441] (II) UnloadModule: "evdev"
[ 25123.447] (II) evdev: Video Bus: Close
[ 25123.447] (II) UnloadModule: "evdev"
[ 25123.450] (II) evdev: Sleep Button: Close
[ 25123.450] (II) UnloadModule: "evdev"
[ 25123.457] (II) evdev: UVC Camera (17ef:4807): Close
[ 25123.457] (II) UnloadModule: "evdev"
[ 25123.457] (II) evdev: AT Translated Set 2 keyboard: Close
[ 25123.457] (II) UnloadModule: "evdev"
[ 25123.464] (II) UnloadModule: "synaptics"
[ 25123.467] (II) evdev: TPPS/2 IBM TrackPoint: Close
[ 25123.467] (II) UnloadModule: "evdev"
[ 25123.467] (II) evdev: ThinkPad Extra Buttons: Close
[ 25123.467] (II) UnloadModule: "evdev"
[ 25123.515] Server terminated successfully (0). Closing log file.

Soo... Xorg.1.log tells me everything died all at once but nothing about why..

cfr · 2012-09-03 23:17:09

There was a thread recently in which somebody claimed that blacklisting btusb magically solved not dissimilar problems. I've also been having issues with bluetooth on a Lenovo machine recently. (I had them initially; a BIOS update sorted them; recently I started having them again.) I'm far from certain that bluetooth has anything to do with it and the person who described the magic solution had no explanation although they certainly took a lot of words not giving one.

If you don't use bluetooth (or, for testing purposes, don't absolutely depend on it), I would consider disabling bluetooth and seeing if it makes a difference. I would even be inclined to disable it in the bios if you have that option although I think blocking the relevant modules from loading is enough.

This probably won't do any good but it cannot hurt anything that I know of. I've just seen a lot of problems with the bluetooth stuff in Linux and I'm starting to think that some of those problems (or different problems) are resurfacing for some people. But I'm currently suspicious of everything - hardware, firmware, software, Linux, non-Linux, the lot - so this doesn't really mean much...

jcg · 2012-09-04 01:08:10

Thanks for the tip cfr, I'll give it a shot. Unfortunately, I can't seem to locate a bluetooth toggle in the W500 bios. I'll report back if the problem persists with the modules blacklisted.

EDIT: I have done the following:

blacklist bluetooth 
blacklist btusb

The bluetooth indicator light is still shining brightly though. Will report if it happens again.

Last edited by jcg (2012-09-04 01:28:00)

cfr · 2012-09-04 02:04:21

You can tell if the modules are loaded using something like

lsmod | grep blue
lsmod | grep btusb

If you get no results, they weren't loaded. Otherwise, you might want to check the wiki instructions on how to do this. I seem to remember it has changed...

jcg · 2012-09-09 20:48:01

This has happened a few more times since I last posted, and I have gained some new (albeit possibly inconsequential) information. Two more pieces to add to the puzzle:

(1) The mechanism by which the screen is blanking appears to be

s2both

One time that the problem occurred, the console spat out something about s2both suspending then immediately trying to un-suspend. I have my laptop configured to do s2both when the lid is closed.

(2) I was messing around with my Raspberry Pi and I have networkmanager set up to forward my WiFi signal through my Ethernet port on the laptop to the R-Pi's Ethernet port. When the problem occurred this last time I was using my R-Pi in just such a manner. Unfortunately I had not yet configured ssh. I am doing that as we speak. However, the internet connection KEPT WORKING even though the laptop was unresponsive and otherwise in a Persistent Vegetative State. I was able to ping all over the place, including the laptop and my LAN, so I have some hope that ssh, once configured, might be a route to recovery. This isn't a solution by any means, but it could be useful for investigative purposes.

Last edited by jcg (2012-09-09 23:15:54)

cfr · 2012-09-09 23:14:53

That sounds like a very promising avenue of enquiry...

jcg · 2012-09-10 01:43:31

I bet with my 'Net I can catch this thing yet.

jcg · 2012-09-15 01:31:01

OK, so I caught it again with my R-Pi, ssh-ed into the W500, poked around in /var/log and really didn't find anything interesting...

However, I have noticed a pattern. It appears that after things have been running for a while (usually Chromium with about 10-20 tabs open, emacs, a terminal doing some things maybe building a package or two) when I mouse over the tint2 nav bar BAM dead.

Also, when mousing over the tint2 nav bar (that is, when it doesn't trigger death) I notice that the little box that pops up to indicate which app you have moused over appears, for lack of a better word, garbled.

Now, why should mousing over the tint2 nav bar cause a botched s2both call to occur? That is the big question, and is completely baffling to me...

EDIT: I also noticed a sort of residual window outline effect when I opened the terminal window in Workspace 1 and switched to Workspace 2,3,etc. I am using xcmpmgr for compositing in openbox. So maybe the problem is in xcmpmgr..?

I'm totally and completely lost on this one...

EDIT2: My laptop has 2 graphics cards. It has an ATI FireGL Mobility V5700 (HD3650... kind of) "discrete GPU" and an Intel 4500MHD "integrated GPU". I usually run the Intel GPU (set in the BIOS) because it uses less power. I'm running the ATI GPU now and I'm not able to make X crash by mousing over the Tint2 nav bar. Also, the compositing looks way better. This hints towards the Intel driver as the culprit. Still no idea why in the hell s2both should be triggered...

Last edited by jcg (2012-09-15 19:32:13)

Arch Linux

#26 2012-02-09 15:37:48

Re: Computer freezing how to troubleshoot?

#27 2012-02-13 12:50:55

Re: Computer freezing how to troubleshoot?

#28 2012-02-28 07:45:46

Re: Computer freezing how to troubleshoot?

#29 2012-03-14 19:30:12

Re: Computer freezing how to troubleshoot?

#30 2012-03-18 12:44:28

Re: Computer freezing how to troubleshoot?

#31 2012-03-18 16:39:18

Re: Computer freezing how to troubleshoot?

#32 2012-03-19 08:42:12

Re: Computer freezing how to troubleshoot?

#33 2012-03-19 13:42:13

Re: Computer freezing how to troubleshoot?

#34 2012-03-20 07:29:06

Re: Computer freezing how to troubleshoot?

#35 2012-03-20 10:48:24

Re: Computer freezing how to troubleshoot?

#36 2012-03-23 08:35:04

Re: Computer freezing how to troubleshoot?

#37 2012-03-23 15:05:58

Re: Computer freezing how to troubleshoot?

#38 2012-09-03 02:41:16

Re: Computer freezing how to troubleshoot?

#39 2012-09-03 06:41:04

Re: Computer freezing how to troubleshoot?

#40 2012-09-03 15:13:39

Re: Computer freezing how to troubleshoot?

#41 2012-09-03 18:19:37

Re: Computer freezing how to troubleshoot?

#42 2012-09-03 19:41:34

Re: Computer freezing how to troubleshoot?

#43 2012-09-03 22:27:26

Re: Computer freezing how to troubleshoot?

#44 2012-09-03 23:17:09

Re: Computer freezing how to troubleshoot?

#45 2012-09-04 01:08:10

Re: Computer freezing how to troubleshoot?

#46 2012-09-04 02:04:21

Re: Computer freezing how to troubleshoot?

#47 2012-09-09 20:48:01

Re: Computer freezing how to troubleshoot?

#48 2012-09-09 23:14:53

Re: Computer freezing how to troubleshoot?

#49 2012-09-10 01:43:31

Re: Computer freezing how to troubleshoot?

#50 2012-09-15 01:31:01

Re: Computer freezing how to troubleshoot?

Board footer