You are not logged in.

#1 2013-10-18 19:58:38

highway
Member
Registered: 2012-07-28
Posts: 81

How to diagnose software / driver / video card issue?

My computer has been freezing up on a regular basis for the past few weeks . . .

I initially assumed it to be a bad stick of RAM, but I was able to reproduce the lockup with each stick individually.

I tried downgrading the kernel (but not Xorg or nvidia drivers), and that didn't fix anything.

At this point I don't really know enough (especially related to Arch specifically) as to how (which) log files to check for a reported error and issue.

The main issue of locking up occurs when I'm watching something on VLC at full screen. It's exacerbated with browsing or doing other stuff at the same time.

After a lockup I don't know if I should be checking journalctl / dmesg / Xorg.log / htop / etc . . .

Any advice / direction of stuff I can be checking after a crash would be MUCH appreciated. It really sucks not being able to use my PC like I used to.

If it help any my board/RAM are ~6yrs old, but the gfx card is new as of this year (I am using the regular nvidia drivers and multilib): EVGA 02G-P4-2645-KR GeForce GT 640

http://www.newegg.com/Product/Product.a … 6814130796

Offline

#2 2013-10-18 21:26:37

ooo
Member
Registered: 2013-04-10
Posts: 1,637

Re: How to diagnose software / driver / video card issue?

did you test your memory with memtest86+ or similar? you can do it with your arch linux boot media, just select memtest86 or whatever from the grub menu.

Offline

#3 2013-10-18 21:29:59

highway
Member
Registered: 2012-07-28
Posts: 81

Re: How to diagnose software / driver / video card issue?

ooo wrote:

did you test your memory with memtest86+ or similar? you can do it with your arch linux boot media, just select memtest86 or whatever from the grub menu.

yup, i ran memtest86+ off of a live cd for 2hrs / 4 passes with no errors . . . I'd like to start looking into the logs to see what comes up after the PC locks from using the video card / watching video, but I don't know where to start checking since I don't really know what I'm looking for or which logging / diagnostic tools I should be using.

Offline

#4 2013-10-18 23:20:04

ooo
Member
Registered: 2013-04-10
Posts: 1,637

Re: How to diagnose software / driver / video card issue?

I would try if the latest nvidia beta driver if it happened to fix this. also have you messed around with any gpu driver settings or similar that could be buggy or unstable during past weeks?

I don't know what happens to any logging services if your system freezes, so I can't help you much there unfortunately, but you could poke around journalctl and anything in /var/log/

Offline

#5 2013-10-18 23:52:02

highway
Member
Registered: 2012-07-28
Posts: 81

Re: How to diagnose software / driver / video card issue?

ooo wrote:

I would try if the latest nvidia beta driver if it happened to fix this. also have you messed around with any gpu driver settings or similar that could be buggy or unstable during past weeks?

I don't know what happens to any logging services if your system freezes, so I can't help you much there unfortunately, but you could poke around journalctl and anything in /var/log/

thanks! i actually don't touch anything . . . everything is all standard install, i like my linux a little custom with a wm and stuff, but i don't mess with the internals of stuff by any means . . .

Offline

#6 2013-10-20 16:39:46

ooo
Member
Registered: 2013-04-10
Posts: 1,637

Re: How to diagnose software / driver / video card issue?

well.. that should be a good approach for stability smile

One more thing you could try is using different video output for vlc. I don't know which one it uses by default, but I would imagine vaapi for example could still be buggy on some setups.

hope you get this sorted out..

EDIT: you could also try if using noveau driver solves your issue. if not, you could at least rule out the nvidia-drivers as a culprit

Last edited by ooo (2013-10-20 20:03:26)

Offline

#7 2013-10-21 02:54:31

highway
Member
Registered: 2012-07-28
Posts: 81

Re: How to diagnose software / driver / video card issue?

ooo wrote:

well.. that should be a good approach for stability smile

One more thing you could try is using different video output for vlc. I don't know which one it uses by default, but I would imagine vaapi for example could still be buggy on some setups.

hope you get this sorted out..

EDIT: you could also try if using noveau driver solves your issue. if not, you could at least rule out the nvidia-drivers as a culprit

doesn't seem too stable now tongue

i'll give the other stuff a try . . . i have nothing left to try at this point

Offline

#8 2013-10-24 21:24:02

Ibex
Member
Registered: 2006-03-02
Posts: 135

Re: How to diagnose software / driver / video card issue?

I'm going to track this topic as I suffer from the same issue the last month or so.

Maybe you can check journalctl? You can go to the last page by pressing > and then using page up and down to navigate to the moment right before the last crash. In my case however, it didn't show anything of interest.

Offline

#9 2013-10-24 21:33:22

highway
Member
Registered: 2012-07-28
Posts: 81

Re: How to diagnose software / driver / video card issue?

Ibex wrote:

I'm going to track this topic as I suffer from the same issue the last month or so.

Maybe you can check journalctl? You can go to the last page by pressing > and then using page up and down to navigate to the moment right before the last crash. In my case however, it didn't show anything of interest.

are you using nvidia with vlc? or do you think something else is causing your lockup?

Offline

#10 2013-10-25 05:13:58

Ibex
Member
Registered: 2006-03-02
Posts: 135

Re: How to diagnose software / driver / video card issue?

highway wrote:
Ibex wrote:

I'm going to track this topic as I suffer from the same issue the last month or so.

Maybe you can check journalctl? You can go to the last page by pressing > and then using page up and down to navigate to the moment right before the last crash. In my case however, it didn't show anything of interest.

are you using nvidia with vlc? or do you think something else is causing your lockup?

No, I use Core i3 integrated graphics on an asus P8B WS motherboard for a headless server. So the cause might be different, but the behavior is the same; weird freezes that can't be pinpointed to a certain cause.

My server reboots after a certain amount of time frozen. Last night, it seems it crashed again, and there's a 3 minute gap in the logging. The night before, it was a 6 minute gap. Does your machine behaves the same?

Offline

#11 2013-10-25 05:19:16

highway
Member
Registered: 2012-07-28
Posts: 81

Re: How to diagnose software / driver / video card issue?

Ibex wrote:
highway wrote:
Ibex wrote:

I'm going to track this topic as I suffer from the same issue the last month or so.

Maybe you can check journalctl? You can go to the last page by pressing > and then using page up and down to navigate to the moment right before the last crash. In my case however, it didn't show anything of interest.

are you using nvidia with vlc? or do you think something else is causing your lockup?

No, I use Core i3 integrated graphics on an asus P8B WS motherboard for a headless server. So the cause might be different, but the behavior is the same; weird freezes that can't be pinpointed to a certain cause.

My server reboots after a certain amount of time frozen. Last night, it seems it crashed again, and there's a 3 minute gap in the logging. The night before, it was a 6 minute gap. Does your machine behaves the same?

yuuuuuuuuuuuuuup, reboots on its own if i don't touch it, but i usually freak out a bit and hit the reset button before it gets a chance to do it on its own . . .

glad to hear it isn't just me, but sucks that it's happening to others . . .

Offline

#12 2013-10-26 09:53:05

Ibex
Member
Registered: 2006-03-02
Posts: 135

Re: How to diagnose software / driver / video card issue?

Ok, I hope there is someone with more knowledge that can help us out, finding what's going on.

Maybe we can check what our systems have in common. Can you show me the output of lspci?

Offline

#13 2013-10-27 15:39:57

highway
Member
Registered: 2012-07-28
Posts: 81

Re: How to diagnose software / driver / video card issue?

Ibex wrote:

Ok, I hope there is someone with more knowledge that can help us out, finding what's going on.

Maybe we can check what our systems have in common. Can you show me the output of lspci?

https://gist.github.com/anonymous/bec7333f781d58c0fe50

Offline

#14 2013-10-28 03:37:37

highway
Member
Registered: 2012-07-28
Posts: 81

Re: How to diagnose software / driver / video card issue?

switched from awesome-wm to e17 and i'm seeing favorable results . . . video with vlc has stuttered heavily several times, and has seemed to lock up / freeze, only to resume playback after a 2-3 second video delay that catches right back up with the audio. i don't know if these are examples of the same problem, but i will post back with further results . . .

i am going to run htop now and see if i can find any events that coincide with the heavy stutters i've seen so far . . .

Offline

#15 2013-11-06 22:34:03

siberianli
Member
Registered: 2013-11-06
Posts: 7

Re: How to diagnose software / driver / video card issue?

Hi, I also have stability issues so I am also interested in tips how to diagnose an unstable system. In my case, in journalctl the problems right before freezing can be seen, e.g.

Nov 05 21:42:59 HT-PC kernel: BUG: unable to handle kernel paging request at ffffff7f811b3f25
Nov 05 21:42:59 HT-PC kernel: IP: [<ffffff7f811b3f25>] 0xffffff7f811b3f24
Nov 05 21:42:59 HT-PC kernel: PGD 0 
Nov 05 21:42:59 HT-PC kernel: Oops: 0010 [#1] PREEMPT SMP 
Nov 05 21:42:59 HT-PC kernel: Modules linked in: fuse nvidia(PO) drm bnep snd_hda_codec_hdmi rc_anysee snd_hda_codec_rea
Nov 05 21:42:59 HT-PC kernel: CPU: 0 PID: 769 Comm: plugin-containe Tainted: P         C O 3.11.6-1-ARCH #1
Nov 05 21:42:59 HT-PC kernel: Hardware name: nVidia MCP7A/To be filled by O.E.M., BIOS 080015  03/24/2010
Nov 05 21:42:59 HT-PC kernel: task: ffff8800ca6cedb0 ti: ffff8800ca17a000 task.ti: ffff8800ca17a000
Nov 05 21:42:59 HT-PC kernel: RIP: 0010:[<ffffff7f811b3f25>]  [<ffffff7f811b3f25>] 0xffffff7f811b3f24
Nov 05 21:42:59 HT-PC kernel: RSP: 0018:ffff8800ca17bf50  EFLAGS: 00010246
Nov 05 21:42:59 HT-PC kernel: RAX: 0000000000000001 RBX: 00000000ffffffff RCX: 00007f55285b3b78
Nov 05 21:42:59 HT-PC kernel: RDX: 0000000000000001 RSI: 0000000000000000 RDI: 0000000000000000
Nov 05 21:42:59 HT-PC kernel: RBP: ffff8800ca17bf78 R08: ffff8800ca17bbac R09: 0000000000000000
Nov 05 21:42:59 HT-PC kernel: R10: 0000000000000001 R11: 0000000000000001 R12: 0000000000000001
Nov 05 21:42:59 HT-PC kernel: R13: 00007f55285b3b70 R14: 0000000000000001 R15: 00007f55285b4700
Nov 05 21:42:59 HT-PC kernel: FS:  00007f55285b4700(0000) GS:ffff88011fc00000(0000) knlGS:0000000000000000
Nov 05 21:42:59 HT-PC kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Nov 05 21:42:59 HT-PC kernel: CR2: ffffff7f811b3f25 CR3: 00000000ca16d000 CR4: 00000000000407f0
Nov 05 21:42:59 HT-PC kernel: Stack:
Nov 05 21:42:59 HT-PC kernel:  00007f5535615320 00007f553550ad78 00007f55285b3b70 0000000000000001
Nov 05 21:42:59 HT-PC kernel:  00007f5536cf1300 00007f55285b3be0 ffffffff814ea5dd 0000000000000293
Nov 05 21:42:59 HT-PC kernel:  0000000000000400 0000000000000000 0000000000000000 0000000000000007
Nov 05 21:42:59 HT-PC kernel: Call Trace:
Nov 05 21:42:59 HT-PC kernel:  [<ffffffff814ea5dd>] system_call_fastpath+0x1a/0x1f
Nov 05 21:42:59 HT-PC kernel: Code:  Bad RIP value.
Nov 05 21:42:59 HT-PC kernel: RIP  [<ffffff7f811b3f25>] 0xffffff7f811b3f24
Nov 05 21:42:59 HT-PC kernel:  RSP <ffff8800ca17bf50>
Nov 05 21:42:59 HT-PC kernel: CR2: ffffff7f811b3f25
Nov 05 21:42:59 HT-PC kernel: ---[ end trace 28ebc192d739a417 ]---

I tried various video drivers; nvidia 304, 325, 331 and nouveau, but in any case the system freezes after playing video for half an hour or so. The problem is more rare on a 32 bit installation than 64 bit, so perhaps you could move from 64 to 32 bit?

Offline

#16 2013-11-08 03:16:42

highway
Member
Registered: 2012-07-28
Posts: 81

Re: How to diagnose software / driver / video card issue?

siberianli wrote:

Hi, I also have stability issues so I am also interested in tips how to diagnose an unstable system. In my case, in journalctl the problems right before freezing can be seen, e.g.

Nov 05 21:42:59 HT-PC kernel: BUG: unable to handle kernel paging request at ffffff7f811b3f25
Nov 05 21:42:59 HT-PC kernel: IP: [<ffffff7f811b3f25>] 0xffffff7f811b3f24
Nov 05 21:42:59 HT-PC kernel: PGD 0 
Nov 05 21:42:59 HT-PC kernel: Oops: 0010 [#1] PREEMPT SMP 
Nov 05 21:42:59 HT-PC kernel: Modules linked in: fuse nvidia(PO) drm bnep snd_hda_codec_hdmi rc_anysee snd_hda_codec_rea
Nov 05 21:42:59 HT-PC kernel: CPU: 0 PID: 769 Comm: plugin-containe Tainted: P         C O 3.11.6-1-ARCH #1
Nov 05 21:42:59 HT-PC kernel: Hardware name: nVidia MCP7A/To be filled by O.E.M., BIOS 080015  03/24/2010
Nov 05 21:42:59 HT-PC kernel: task: ffff8800ca6cedb0 ti: ffff8800ca17a000 task.ti: ffff8800ca17a000
Nov 05 21:42:59 HT-PC kernel: RIP: 0010:[<ffffff7f811b3f25>]  [<ffffff7f811b3f25>] 0xffffff7f811b3f24
Nov 05 21:42:59 HT-PC kernel: RSP: 0018:ffff8800ca17bf50  EFLAGS: 00010246
Nov 05 21:42:59 HT-PC kernel: RAX: 0000000000000001 RBX: 00000000ffffffff RCX: 00007f55285b3b78
Nov 05 21:42:59 HT-PC kernel: RDX: 0000000000000001 RSI: 0000000000000000 RDI: 0000000000000000
Nov 05 21:42:59 HT-PC kernel: RBP: ffff8800ca17bf78 R08: ffff8800ca17bbac R09: 0000000000000000
Nov 05 21:42:59 HT-PC kernel: R10: 0000000000000001 R11: 0000000000000001 R12: 0000000000000001
Nov 05 21:42:59 HT-PC kernel: R13: 00007f55285b3b70 R14: 0000000000000001 R15: 00007f55285b4700
Nov 05 21:42:59 HT-PC kernel: FS:  00007f55285b4700(0000) GS:ffff88011fc00000(0000) knlGS:0000000000000000
Nov 05 21:42:59 HT-PC kernel: CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
Nov 05 21:42:59 HT-PC kernel: CR2: ffffff7f811b3f25 CR3: 00000000ca16d000 CR4: 00000000000407f0
Nov 05 21:42:59 HT-PC kernel: Stack:
Nov 05 21:42:59 HT-PC kernel:  00007f5535615320 00007f553550ad78 00007f55285b3b70 0000000000000001
Nov 05 21:42:59 HT-PC kernel:  00007f5536cf1300 00007f55285b3be0 ffffffff814ea5dd 0000000000000293
Nov 05 21:42:59 HT-PC kernel:  0000000000000400 0000000000000000 0000000000000000 0000000000000007
Nov 05 21:42:59 HT-PC kernel: Call Trace:
Nov 05 21:42:59 HT-PC kernel:  [<ffffffff814ea5dd>] system_call_fastpath+0x1a/0x1f
Nov 05 21:42:59 HT-PC kernel: Code:  Bad RIP value.
Nov 05 21:42:59 HT-PC kernel: RIP  [<ffffff7f811b3f25>] 0xffffff7f811b3f24
Nov 05 21:42:59 HT-PC kernel:  RSP <ffff8800ca17bf50>
Nov 05 21:42:59 HT-PC kernel: CR2: ffffff7f811b3f25
Nov 05 21:42:59 HT-PC kernel: ---[ end trace 28ebc192d739a417 ]---

I tried various video drivers; nvidia 304, 325, 331 and nouveau, but in any case the system freezes after playing video for half an hour or so. The problem is more rare on a 32 bit installation than 64 bit, so perhaps you could move from 64 to 32 bit?

sorry for your troubles! i am not sure what the exact issue is, but i am on nvidia w/ video troubles as well . . . i have been running atop 24/7 and notice high r/w use of my disk at times when playback starts to stutter . . .

i haven't had a full-on lockup in days, but have still had one in recent time. i am not really sure how to trouble shoot this further.

Offline

#17 2013-11-08 11:27:59

mr.MikyMaus
Member
From: disabled
Registered: 2006-03-31
Posts: 285

Re: How to diagnose software / driver / video card issue?

Did you try to clean heatsinks and friends?


What happened to Arch's KISS? systemd sure is stupid but I must have missed the simple part ...

... and who is general Failure and why is he reading my harddisk?

Offline

#18 2013-11-10 20:01:05

siberianli
Member
Registered: 2013-11-06
Posts: 7

Re: How to diagnose software / driver / video card issue?

Good suggestion to look at heating problems. With nvidia-settings -q gpucoretemp the gpu temperature is checked during playing video under my 64 bit arch system. At 43 C video freezes but strange enough, sound continues to play. The mouse can't be moved but it is possible to remotely log in with ssh. I reboot into the 32 bit installation (I have a dual boot available with both 64 bit and 32 bit) and when the video is continued the temperature slowly increases up to 47 C but the system doesn't freeze. So I rule out that heating causes freezing the system.

Offline

#19 2013-11-10 22:32:49

MW
Member
Registered: 2007-07-27
Posts: 127

Re: How to diagnose software / driver / video card issue?

With a PC freezing at random, try it with a LiveCD. Put some stress on the GPU and CPU. This will either show you it is hardware or software related. If it still freezes then it is hardware related. If nothing happens, then it is probably software related.

Hardware causes, I usually start with
1. Faulty PSU
2. Overheating
3. CPU heat sink seated properly?
4. Faulty memory
5. Faulty on-board controller (SATA, etc.)
6. Failing HDD

Software causes, check the logs, there should be some indication as to what happened and why the PC froze. Can you reproduce it in another WM or DE? Can you reproduce it with vesa drivers? What happens if you boot an older kernel or even a fail-safe kernel?
Is there a pattern? The PC locks up after a certain amount of uptime, after a game I played or program I used, after it came back from hibernation, etc. All this will help narrow down what went wrong.

Offline

#20 2013-11-13 22:34:22

siberianli
Member
Registered: 2013-11-06
Posts: 7

Re: How to diagnose software / driver / video card issue?

All right, let's try a LiveCD then. The arch installation CD is unsuitable since it doesn't include a graphical environment, but ArchBang offers a LiveCD that I happen to have at home. But unfortunately, after loading the kernel and starting services like networking, a kernel panic occurs and I can't proceed. I get myself an Ubuntu 13.10 64 bit LiveCD that boots into the desktop. Stress testing is difficult as all video requires MPEG plugins that are missing in the LiveCD environment. However, even without any stress dmesg shows several times 'BUG: unable to handle kernel paging request  at fffffff' and on trying to start firefox the system seems to hang (the pointer changes to the wait cursor, it can be moved but there is no response to mouse clicks).

So this points in the direction of a hardware problem. But what if my PC freezes on one LiveCD but works fine on another LiveCD, would that still mean that the hardware is faulty? I will try the 32 bit version of the Ubuntu LiveCD to see if the same problems occur, I suspect they don't. As for hardware causes, I checked GPU temp and memory, but am not sure how to check most of the others, other than replacing parts.

Offline

#21 2013-11-14 21:03:14

siberianli
Member
Registered: 2013-11-06
Posts: 7

Re: How to diagnose software / driver / video card issue?

Today it is time for a test of a newOS on a LiveCD (one OS a day keeps the doctor away as the saying goes); Ubuntu 12.04 32 bit. It is possible to start firefox and install a flash plugin and now I have played two youtube videos for about 60 mins without any interrupts. I notice in dmesg some kernel page errors, but these problems do not freeze the system.

I like to come back to my previous question; what does it mean that the system freezes in one OS and doesn't in another? I suspect that my mother board pretends to support 64 bit but that it is a bluff and quickly panics when we actually try it. Does this make sense? Or there is a problem both in 32 bit and 64 bit but the 64 bit OS is more sensitive?

Offline

#22 2013-11-18 19:49:07

siberianli
Member
Registered: 2013-11-06
Posts: 7

Re: How to diagnose software / driver / video card issue?

My motherboard is a Zotac ION GF9300 with integrated GPU and I am tempted to add a low budget graphics card to the system to try solve the instability problems, e.g. a 25 euro EVGA GeForce 210.

Last edited by siberianli (2013-11-18 19:50:12)

Offline

#23 2013-12-03 03:19:10

highway
Member
Registered: 2012-07-28
Posts: 81

Re: How to diagnose software / driver / video card issue?

siberianli wrote:

My motherboard is a Zotac ION GF9300 with integrated GPU and I am tempted to add a low budget graphics card to the system to try solve the instability problems, e.g. a 25 euro EVGA GeForce 210.

not sure about the gfx card fixing things, i guess it's worth a shot . . . usually one OS and not another means that it's that particular OS and not a hardware issue.

Offline

#24 2014-01-04 08:04:18

siberianli
Member
Registered: 2013-11-06
Posts: 7

Re: How to diagnose software / driver / video card issue?

Sorry for my late answer, but I gave the addition of a gfx card a shot as mentioned above and this did not change anything. Since my system has an integrated GPU I checked that the newly added card was indeed the one that was being used. It is currently hard to conclude; the instability issues depend on the OS (pointing at an OS issue) but occur in multiple OS-es and multiple drivers (pointing at a hardware issue). Arch 32 bit is relatively stable and 64 bit versions of arch and ubuntu are unstable.

Last edited by siberianli (2014-01-04 12:10:56)

Offline

#25 2014-01-07 19:46:50

siberianli
Member
Registered: 2013-11-06
Posts: 7

Re: How to diagnose software / driver / video card issue?

Reading back on this thread I realize that my issues seem unrelated to those of highway and Ibex since I see kernel paging errors in my log files and they don't. I guess it is better to start a new thread for that...

Offline

Board footer

Powered by FluxBB