Nvidia kernel issue

[Jeff]

I’m having a consistent problem where after some period of inactivity (sometimes just long enough for the screen to go black, sometimes hours), the nvidia drivers appear to have a problem and then cause Gnome / X to restart, killing all graphical applications, and pretty much ruining my day.

For a while I could prevent this from happening by disabling autoupdate. However, now even that doesn’t keep my system stable.

I’m sure I don’t have to say how frustrating this is. Almost as frustrating as running the noveau driver, so I switch back and forth depending on how much risk I’m willing to take.

In looking at journalctl I’m seeing the following line right before the trouble:
nvidia-modeset: ERROR: GPU:0: Failed to allocate memory for composition pipeline

I installed the nvidia drivers using the scripts linked to here:

Bash scripts to automate installation of NVIDIA proprietary driver

Available here Table of Contents Summary Detailed description for each script pre_install.bash install.bash post_install.bash uninstall.bash pre_update.bash update.bash post_update.bash cleaning.bash Known Issues Summary Based on the tutorial by Clear Linux, I wrote several bash scripts that automate the installation, update, and uninstallation, of NVIDIA proprietary driver. There’s no need to give them execute permission to make them executable. To run a script, execute bash ./<SCR…

I generally have no problem running the scripts to install or uninstall the drivers, and while using the machine, I don’t experience issues. Its only when I step away. Sometimes it can be days in between when it happens, and sometimes its under an hour.

And help in further debugging and resolving this would be greatly appreciated.

[lemon-bean]

I’ve got the same problem. I use RTX 2060 to do some Deep Learning computation almost every day for about a month and it fails to wake up some times from the screen turning off since yesterday.

[doct0rHu]

I wish this is not cause by the recent kernel update…

[aluo-x]

I also have the same issue with Nvidia 1080TIs. After inactivity (screen goes dark), the computer can no longer be interacted with via mouse. This only happens sometimes, so I haven’t figured out what causes it.

[doct0rHu]

Well this doesn’t happen to me. I’m on latest kernel.
You can try the LTS kernel.

[Ryan_Murphy]

Mine crashes all the time to especially when lefft sitting. I just switched RAM out and filled every slot with same RAM and it seems to be more stable. I read somewhere of a regression with 5.9 kernel and this is probably what they were talking about. Maybe 5.10 will be all good.

The Latest On The Linux 5.9 Kernel Regression Stemming From Page Lock Fairness