[nvidia] gnome-shell GUI freezes totally

Bug #1887821 reported by Sami Kuisma
12
This bug affects 2 people
Affects Status Importance Assigned to Milestone
Ubuntu
Undecided
Unassigned

Bug Description

I'm using basic Ubuntu 20.04 and GUI (gnome?) freezes totally.
- I was running Firefox 78.0.2 and watching Youtube. The video and sound stopped.
- Keyboard was unresponsive, can't get into console, and for example pressing caps lock/num lock didn't toggle keyboard leds.
- mouse didn't move.
- The whole computer and OS didn't crash I suppose. I could ping the computer from another computer. and there were entries in /var/log/syslog after the GUI freeze.

What ever happened I can find some indicating in /var/log/syslog:

Crash happened 14:21 local time which pinpoints the bug into these entries:

Jul 16 14:21:24 Aorus-Ultra-Ryzen gnome-shell[897367]: [Child 897367, ImageBridgeChld] WARNING: failed to open shm: Too many open files: file /build/firefox-k3d8Rk/firefox-78.0.2+build2/ipc/chromium/src/base/shared_memory_posix.cc, line 246
Jul 16 14:21:24 Aorus-Ultra-Ryzen gnome-shell[897367]: [Child 897367, MediaDecoderStateMachine #1] WARNING: Decoder=7f08246e5000 Decode error: NS_ERROR_OUT_OF_MEMORY (0x8007000e) - mozilla::MediaResult mozilla::FFmpegVideoDecoder<46465650>::CreateImage(int64_t, int64_t, int64_t, MediaDataDecoder::DecodedData &): image allocation error: file /build/firefox-k3d8Rk/firefox-78.0.2+build2/dom/media/MediaDecoderStateMachine.cpp, line 3470
Jul 16 14:21:25 Aorus-Ultra-Ryzen gnome-shell[897367]: message repeated 2 times: [ [Child 897367, MediaDecoderStateMachine #1] WARNING: Decoder=7f08246e5000 Decode error: NS_ERROR_OUT_OF_MEMORY (0x8007000e) - mozilla::MediaResult mozilla::FFmpegVideoDecoder<46465650>::CreateImage(int64_t, int64_t, int64_t, MediaDataDecoder::DecodedData &): image allocation error: file /build/firefox-k3d8Rk/firefox-78.0.2+build2/dom/media/MediaDecoderStateMachine.cpp, line 3470]

after that anything GUI, keyboard or mouse related was totally unresponsive.

there were other entries in syslog after that, and ping worked into this frozen machine. So the whole OS didn't crash. Only GUI related.

I hope this helps.

lsb_release -rd
Description: Ubuntu 20.04 LTS
Release: 20.04

apt-cache policy firefox
firefox:
  Installed: 78.0.2+build2-0ubuntu0.20.04.1
  Candidate: 78.0.2+build2-0ubuntu0.20.04.1
  Version table:
 *** 78.0.2+build2-0ubuntu0.20.04.1 500
        500 http://fi.archive.ubuntu.com/ubuntu focal-updates/main amd64 Packages
        500 http://security.ubuntu.com/ubuntu focal-security/main amd64 Packages
        100 /var/lib/dpkg/status
     75.0+build3-0ubuntu1 500
        500 http://fi.archive.ubuntu.com/ubuntu focal/main amd64 Packages

I'm sure this is a bug. The same effect has happened 3 times in 30 days. I didn't check the previous syslog which error message there was.

Sami Kuisma (sku-8936)
affects: pulseaudio (Ubuntu) → gnome-shell (Ubuntu)
Revision history for this message
Sami Kuisma (sku-8936) wrote :

System:

- Gigabyte X570 Aorus Ultra
Gigabyte Technology Co., Ltd. X570 AORUS ULTRA/X570 AORUS ULTRA, BIOS F3 06/13/2019

- AMD Ryzen 5 3600
- 64GB ram

- gfx: GTX 680
NVIDIA-SMI 440.95.01 Driver Version: 440.95.01 CUDA Version: 10.2

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Please:

1. Run this command to send us more info about the machine:

   apport-collect 1887821

2. Next time the problem happens and you have to reboot, please then run:

   journalctl -b-1 > prevboot.txt

   and attach the resulting text file here.

3. Check for crashes by following *all* of these instructions:

   https://wiki.ubuntu.com/Bugs/Responses#Missing_a_crash_report_or_having_a_.crash_attachment

tags: added: focal
Changed in gnome-shell (Ubuntu):
status: New → Incomplete
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Also...

4. Run this command:

   lspci -k > lspcik.txt

   and attach the resulting text file here.

5. If the problem is just the machine running out of memory then please track that using the 'free -h' and 'ps auwx' commands and find out which process is using the most memory.

affects: gnome-shell (Ubuntu) → ubuntu
Revision history for this message
Sami Kuisma (sku-8936) wrote :

ok, here is prevboot.txt

it was too big, over 8M lines, because of one buggy sw; Interactive Brokers TradeStation floods the log.

wc prevboot.txt
  5089905 91689782 962690406 prevboot.txt

with entries like this:

head prevboot.txt
-- Logs begin at Tue 2020-07-14 17:11:55 EEST, end at Fri 2020-07-17 07:38:59 EEST. --
heinä 14 17:11:55 Aorus-Ultra-Ryzen install4j_17tjtje-tws.desktop[793522]: JTS-usfarmDispatcherS10-129S10-130: Couldn't write to log file - java.io.IOException: write beyond end of stream
heinä 14 17:11:55 Aorus-Ultra-Ryzen install4j_17tjtje-tws.desktop[793522]: JTS-usfarmDispatcherS10-129S10-130: Couldn't write to log file - java.io.IOException: write beyond end of stream
heinä 14 17:11:55 Aorus-Ultra-Ryzen install4j_17tjtje-tws.desktop[793522]: JTS-usfarmDispatcherS10-129S10-130: Couldn't write to log file - java.io.IOException: write beyond end of stream

So I removed those tradestation lines and resulting prevboot is 8203 lines

Revision history for this message
Sami Kuisma (sku-8936) wrote :

lspci -k > lspcik.txt

Revision history for this message
Daniel van Vugt (vanvugt) wrote :

The log in comment #4 is missing the time from when the machine booted, which I also wanted to see, but maybe that doesn't matter... Near the end of the log it seems your system is running out of file descriptors AND running out of memory. So I suggest you dig deeper to find out which process(es) are using all the memory and files. Some commands that may help:

  free -h
  ps auxw
  lsof

tags: added: nvidia
summary: - gnome-shell GUI freezes totally
+ [nvidia] gnome-shell GUI freezes totally
Revision history for this message
Sami Kuisma (sku-8936) wrote :

"The log in comment #4 is missing the time from when the machine booted, which I also wanted to see, but maybe that doesn't matter..."

the uptime was 16d and system didn't have logs from the boot time.
the log had everything created by 'journalctl -b-1 > prevboot.txt' except grep -v 'install4j_17tjtje-tws.desktop' lines

The system hadn't been running low on memory for days, unless it didn't happen very suddenly because of some bug in firefox etc. I have 64GB and using roughly 20GB for days (I have top/htop running in terminal all the time). I am typically aware of mem usage.

Revision history for this message
Sami Kuisma (sku-8936) wrote :

as far as I know that CRASH file wasn't created.

- nothing in /var/crash
- nothing recent in https://errors.ubuntu.com/user/e003a24bbaa04f1c5883029a977e36cc1f8f31ebf57b3f3df099cf9cb7d38bb43d671c6a7a2a6376d5385e8c7cbfe0f060bada6610fa542eb5f571a1fadd9fee

if the GUI process hadn't yet crashed, just frozen. as I couldn't get into console (keyboard didn't work) I pressed the reset button in computer. And there wasn't any crash file existing.

Now I installed ssh server, and if the gui freeze happens next time, I can log in from another computer and see processes which is hung, or hogging CPU or mem. The GUI freeze has happened 3 times since initial OS install about a month ago.

Revision history for this message
Sami Kuisma (sku-8936) wrote :

Same sort of total freeze again. GUI totally frozen, while watching youtube video in firefox.

Revision history for this message
Sami Kuisma (sku-8936) wrote :
Revision history for this message
Sami Kuisma (sku-8936) wrote :

frozen before timestamp Sep 2 21:07:30 (which was next reboot)

Revision history for this message
Sami Kuisma (sku-8936) wrote :

could the freeze be something GPU related?

Sep 2 21:04:51 Aorus-Ultra-Ryzen kernel: [935088.204347] NVRM: GPU at PCI:0000:09:00: GPU-1aac77c3-3e8f-aca5-1c8e-19e660d74551
Sep 2 21:04:51 Aorus-Ultra-Ryzen kernel: [935088.204351] NVRM: Xid (PCI:0000:09:00): 62, pid=826, 0c83(1780) 00000000 00000000

the system froze around 21:05 (next reboot was 21:07:30)

Revision history for this message
Sami Kuisma (sku-8936) wrote :
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

It's suspicious the problem in comment #10 occurred as the Nvidia driver was starting. Also that seems to be an odd version of the Nvidia driver we don't distribute.

If you installed the Nvidia driver manually then please uninstall it and then use the 'Additional Drivers' app to install a supported version.

If you installed the Nvidia driver from an older Ubuntu update then your system is out of date and needs updating now. So run:

  sudo apt update
  sudo apt full-upgrade
  Reboot.

Revision history for this message
Sami Kuisma (sku-8936) wrote :

same sort of crash again.... while watching youtube on firefox

might again refer to GPU

last lines in log:

syys 05 17:21:17 Aorus-Ultra-Ryzen kernel: NVRM: GPU at PCI:0000:09:00: GPU-1aac77c3-3e8f-aca5-1c8e-19e660d74551
syys 05 17:21:17 Aorus-Ultra-Ryzen kernel: NVRM: Xid (PCI:0000:09:00): 62, pid=744, 0c83(1780) 00000000 00000000

Revision history for this message
Sami Kuisma (sku-8936) wrote :
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

The messages you refer to are not a crash. And you still have an unsupported version of the Nvidia driver installed. Please follow the instructions in comment #14.

Revision history for this message
Launchpad Janitor (janitor) wrote :

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

Changed in ubuntu:
status: Incomplete → Expired
Revision history for this message
hereiam (jenni-luder) wrote :

I have the same issue on my PC. The gui expected the mouse freezes randomly.

I use Gefore GTX 1050 Ti (GP107) and teh driver version 460 (proprietary, tested)

hereiam (jenni-luder)
Changed in ubuntu:
status: Expired → New
Revision history for this message
Daniel van Vugt (vanvugt) wrote :

Please don't reopen other peoples' expired bugs. You should open your own new bug instead.

Changed in ubuntu:
status: New → Invalid
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers