Every article tag can be clicked to get a list of all articles in that category. Every article tag also has an RSS feed! You can customize an RSS feed too!
Random crashes
Page: 1/3»
  Go to:
PTLyon Dec 20, 2016
Hi guys.

I don't know where else to go. Tried Ubuntu and Mint forums, but no one could help me. We are all linux gamers here, so maybe you can help me, idk.

The problem:
For quite some weeks I've been dealing with random crashes. The computer would just freeze, I wasn't able to use alt+f2 or get any response, the mouse didn't respond. Sometimes it would moves for 10 seconds or so, and then it would freeze, too. But if if there was music playing, it would keep playing. I had to hard reset the computer.
I couldn't isolate a cause for the freeze. It could happen when I'm playing a game (after 1 hour or after 5 minutes), or when I was just browsing the web, or doing nothing at all. It seemed random. Some times two crashes a day, sometimes 8 a day, if I was really lucky, no crashes at all.
I was using Ubuntu 16.04 with nvidia drivers from the repositories.

At first I thought it was hardware. I was about to buy a GTX 960 anyway, so I did, and swapped my gpu. But the problem continued.

Then I took the dust off my dual-booted Windows, and started playing there. It was working fine. So there wasn't any hardware problem.
The only difference is that Windows is on a old HDD, and Ubuntu in a 6 months SSD. But this SSD passes all SMART tests with maximum grade, so I don't think the problem is there.

Although I'm a linux user for some years now, I'm not an expert, I'm just a normal user. So I decided to try OpenSuse. Same problem. Then Linux Mint, same problem.

(the computer just crashed. Thank God Mozilla is able to recover this text I was writing)

Eventually I become suspicious of Nvidia drivers. I was using the ones from the repositories, it was probably the same version (357.67) on the 3 distros (on Ubuntu and Mint, it is). So I asked for help, and was able to update for the version 375.26, which I'm using now. The computer was ok for a couple of days (maybe a placebo effect? Not sure), but now it's back to the crashes. 4 or 5 today.

I don't know for sure if nvidia drivers are to blame, in here. My GTX 960 is a very popular card, if there was something very wrong with nvidia drivers I'm sure a lot more people would have noticed, and nvidia too. And this happens with both older stable version available on the repositories and with the brand new one.

This is a log I believe from a Ubuntu crash:
Dec  6 17:49:03 eduardo-desktop kernel: [ 8659.726052] NVRM: GPU at PCI:0000:01:00: GPU-c4a80165-86e4-bfbb-ecfa-1d81a70b138a
Dec  6 17:49:03 eduardo-desktop kernel: [ 8659.726062] NVRM: Xid (PCI:0000:01:00): 79, GPU has fallen off the bus.
Dec  6 17:49:03 eduardo-desktop kernel: [ 8659.726062] 
Dec  6 17:49:03 eduardo-desktop kernel: [ 8659.726067] NVRM: GPU at 0000:01:00.0 has fallen off the bus.
Dec  6 17:49:03 eduardo-desktop kernel: [ 8659.726078] NVRM: A GPU crash dump has been created. If possible, please run
Dec  6 17:49:03 eduardo-desktop kernel: [ 8659.726078] NVRM: nvidia-bug-report.sh as root to collect this data before
Dec  6 17:49:03 eduardo-desktop kernel: [ 8659.726078] NVRM: the NVIDIA kernel module is unloaded.
Dec  6 17:49:04 eduardo-desktop gnome-session[1771]: [2016-12-06T17:49:04] [ERR] nvctrl: Failed to retrieve measure of type 10201 for NVIDIA GPU 0
Dec  6 17:49:04 eduardo-desktop gnome-session[1771]: [2016-12-06T17:49:04] [ERR] nvctrl: Failed to retrieve measure of type 50204 for NVIDIA GPU 0
Dec  6 17:49:04 eduardo-desktop gnome-session[1771]: [2016-12-06T17:49:04] [ERR] nvctrl: Failed to retrieve measure of type 90204 for NVIDIA GPU 0
Dec  6 17:49:04 eduardo-desktop gnome-session[1771]: [2016-12-06T17:49:04] [ERR] nvctrl: Failed to retrieve measure of type 210204 for NVIDIA GPU 0
Dec  6 17:49:04 eduardo-desktop gnome-session[1771]: [2016-12-06T17:49:04] [ERR] nvctrl: Failed to retrieve measure of type 110204 for NVIDIA GPU 0
Dec  6 17:49:04 eduardo-desktop gnome-session[1771]: [2016-12-06T17:49:04] [ERR] nvctrl: Failed to retrieve measure of type 20202 for NVIDIA GPU 0
Dec  6 17:49:04 eduardo-desktop gnome-session[1771]: [2016-12-06T17:49:04] [ERR] nvctrl: Failed to retrieve measure of type 20204 for NVIDIA GPU 0


Also, when I use nouveau, on Mint, this appears:



ANY suggestion would be appreciated.

Thank you
Liam Dawe Dec 21, 2016
Have you got an alternate GPU and RAM to test?
Samsai Dec 21, 2016
Are you entirely sure the Windows install was stable? That crash log would indicate a GPU or a motherboard problem of some description. Testing the RAM is also a good idea.
FredO Dec 21, 2016
Have you also tried installing the GPU into a different slot (providing you have a spare one)?
Xpander Dec 21, 2016
it sounds like RAM issue. is your PSU enough for your system?
wolfyrion Dec 21, 2016
There are a lot of things to consider and many of you mention some things to check it out but most of the times the problem isnt a specific hardware but all the components in general.
Most people thing if I dont mess with my PC I will just put everything inside the motherboard and everything will work just fine!
Sure it may works fine, and Windows may boot and work ok but Linux is very sensitive on such things thats why it freezes :P
This thing is called Underclocking or Overclocking
This freezing problem occurs a lot when you have to setup a PC with 32GB RAM or 64GB RAM that works @2000-2400Mhz
If you just plug it in your motherboard and dont do the necessary tweaks on your bios dont expect them to work.
As an IT and a support guy I had many issues because of this and most of the times on Linux operating Systems.

Some tips....

1. Check with your memory manufacturer the correct timings and voltage for your RAM. That goes for your CPU as well.
2. if your freezing is happening most of the times in Rocket League I think is something with the camera setting.
You have to disable either INVERT SWIVEL PITCH or CAMERA SHAKE something like that dont remember exactly.
3. Dont afraid to overclock a bit your PC :P

Good luck! :)
PTLyon Dec 21, 2016
Hi guys, thanks for you help!

I'll answer the questions/sugestions, in order:

- I don't have extra ram to test. I can however test 1 ram card at the time, because I have 2x 4GB.
- I have my old GPU (Nvidia GT 730), but this problem started with it.
- I'll pick something to play on Windows (this weekend will be hard because of Christmas, but after it) to confirm, but I completed Ryse: Son of Rome with no computer crashes, 1 week ago or so. The game would sometimes crash (3 or 4 overall) and Windows displayed a message about the graphic driver stop running, but the computer/OS would not crash, just the game. But then again, there were lots of people complaining about this on Steam reviews of this game, so I believe this is the game's fault. But I'll try playing something else, in any case.
- I only have 1 GPU slot.
- PSU. I no nothing about hardware, and even less about PSU. Some people on IRC suggested that the problem might be power related, because windows and linux work differently regarding power management.
This is my computer:

System:    Host: eduardo-desktop Kernel: 4.4.0-21-generic x86_64 (64 bit)
           Desktop: Cinnamon 3.0.6  Distro: Linux Mint 18 Sarah
Machine:   Mobo: ASRock model: FM2A88M-HD+
           Bios: American Megatrends v: P2.60 date: 08/01/2014
CPU:       Quad core AMD A10-7850K Radeon R7 12 Compute Cores 4C+8G (-MCP-)
           speed/max: 1700/3700 MHz
Graphics:  Card: NVIDIA GM206 [GeForce GTX 960]
           Display Server: X.Org 1.18.3 drivers: nvidia (unloaded: fbdev,vesa,nouveau)
           Resolution: [email protected]
           GLX Renderer: GeForce GTX 960/PCIe/SSE2
           GLX Version: 4.5.0 NVIDIA 375.26
Network:   Card: Realtek RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller
           driver: r8169
Drives:    HDD Total Size: 1250.3GB (11.1% used)
Info:      Processes: 178 Uptime: 19 min Memory: 1087.4/7905.7MB
           Client: Shell (bash) inxi: 2.2.35 



Motherboard: ASRock Intel H81 SK1150
PSU: Nox Urano SX 500w

Now, my PSU didn't have the power cable for this GPU's 6 pin entry. But the guy that built my computer 2 years ago told me to use an adapter, so I have 2 molex cables + an adapter that allows me to power the GPU using it's 6 pin entry. (Gembird 2 x Molex p/ PCI-e 6 Pin)

- A local computer shop isn't exactly an option I like because they'll probably just blame Linux. It wouldn't the first time. Some years ago some guy at the store told me my old laptop's integrated gpu was dead because I was using linux -_-'
- I'll also play something on Linux and keep an eye on the temperature (nvidia settings), and will let you know. With just the browser opened, 37º C.
- It was not crashing on Rocket League or at the specific game ;)
- RAMs are: Kingston DDR3 4096MB 1600Mhz x2 . How do I check voltage and timings?

Again, thank you for the help.
wolfyrion Dec 22, 2016
500w I find it extremely low for your PC

And here goes another story from tech support :P

Windows Setup with 700w PSU and as far as I remember an old Radeon VGA card that required a lot of PSU power but 700w were more than enough for that card.That guy had also a creative sound card with external controller that required as well additional extra power but according to our standards 700w PSU was more than enough.The PSU was a good brand as well but not a high end PSU.

The problem:
While he was playing games out of sudden we were hearing a crackling sound and the whole computer was freezing, sometimes shutting down or blue screen.

We thought it was the sound card because when we were removing the sound card everything was working fine and when we were inserting the card back we were having the problem.
We have tried many things even called creative but no luck on replacing the card or anything else.
So after a week we decided to test the card in another computer and the card was working flawlessly ...
Inserting the card back to the original computer = problems again! Jeezzzzz WTF!!!!?!?!?

We have noticed that the other computer we tested had 1k PSU Thermaltake so we did change the power supply with a thermaltake one and guess what... EVERYTHING WAS WORKING FINE!!!

On my PC with a GTX 980 I have a 1200 CORSAIR PSU, a lot of people find it extreme but considering the needs I have , around 15/20 USB Devices, 6 Hard Disks, 2 DVDS, 32GB of RAM @ 2k and so on I am just happy that everything is working smoothly. :)
PTLyon Dec 24, 2016
Hi guys,


Some temperature reports from yesterday:

Firefox only - 42ºC
Rocket League - 59-61ºC (after 2 complete games)
Door Kickers - 56ºC (this one seems to be very low demanding on resources, though)


It's funny that I had no crashes when playing, yesterday. But today I've only used the browser, and it crashed like 5 times already, it's pretty much unusable. It's on 36ºC - 38ºC now, and I'm pretty sure it's crashing when it's cold too, I don't think the temperatures are the problem.

wolfyrion, thanks for sharing your story. The power related problems could explain it... But I don't have much stuff connected to the PC. A usb mouse, a usb keyboard, a sound jack soundsystem that uses external power supply, and sometimes an usb headset.

After reading your comment, I removed the SATA cable from the Windows HDD and I'm entering directly on the Linux SSD, but the problem remains. I only have some front fan, a SSD, GPU, CPU and the motherboard. It can't be that... can it?
PTLyon Dec 24, 2016
F*ck me. I can't make any sense out of this.

So for the first few hours today, the system was totaly unstable. "Unusable", I said, in the post above. And all I did was surfing with firefox.

After the last reset, I reconnected the Windows driver, was about to go to Windows to test it, but decided to try to run some games on Linux one more time and push it a little bit.

I opened 4 or 5 firefox tabs (youtube included, in auto-play), and I run Rocket League. 2 games, system stable. Then I tried Bioshock Infinite. That first scene is in the sea, with very moving waters, rain, etc. No problem. I changed graphic settings to high quality. No problems at all, the system was stable. Temperatures were 50ºC - 54º C (although I had my PC case opened, to keep an eye on the GPU fans - which seemed fine, they started working as soon as the game started). Kept playing for some more time, and I'm still here. 1 hour or more and no freezes! After having freezed A LOT today. When I wrote the post above, it was not holding even for 10 minutes.

So, it freezes because it's... too cold? Arrrgghh
I'd hate to work on a PC clinic. Those machines sometimes are demoniac!
Xpander Dec 24, 2016
i still think its either ram or psu issue. ram issue sounds more likely cause of the "random" nature of this.

check your dmesg, journalctl -b logs to see whats up, maybe there are hints somewhere.
While you're here, please consider supporting GamingOnLinux on:

Reward Tiers: Patreon. Plain Donations: PayPal.

This ensures all of our main content remains totally free for everyone! Patreon supporters can also remove all adverts and sponsors! Supporting us helps bring good, fresh content. Without your continued support, we simply could not continue!

You can find even more ways to support us on this dedicated page any time. If you already are, thank you!
Login / Register


Or login with...
Sign in with Steam Sign in with Google
Social logins require cookies to stay logged in.