Fedora 33 will not boot after installing NVIDIA driver

Graphics / Linux Linux
,
1

I’m running a fresh install of Fedora 33,
trying to use the NVIDIA driver with a GTX 970
made sure it was up to date with dnf update

When I search lspci it recognizes the card.

I install the akmod-nvidia package, wait ~5 minutes, then run ‘modinfo -F version nvidia’ which returns the expected 455 driver version.

however when I reboot the machine, I get as far as selecting what boot option (choice of a couple different kernels or a recovery mode) and then it hangs at either a black screen with a bkinking curor, or a grey screen with 3 dots and doe not advance even after letting it sit overnight, the machine is also inaccesible by SSH.

I removed the blacklist lines for nouveau and the machine boots, and then ran the Nvidia_bug_report.sh script, the results are at
https://paste.centos.org/view/3aeea05c
I’m kind of at a loss here, not sure if there’s just a step I missed or something I’m misunderstanding.

nvidia-bug-report.log.gz (75.9 KB)

2

Your GPU has fallen off the bus


      journalctl -b -1:
    Oct 29 12:18:42 localhost.localdomain kernel: Command line: BOOT_IMAGE=(hd0,msdos1)/vmlinuz-5.8.16-300.fc33.x86_64 root=UUID=b0f449d5-044c-4de0-a6de-02768c4dbd6a ro rootflags=subvol=root rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau nvidia-drm.modeset=1
    Oct 29 12:18:42 localhost.localdomain kernel: Kernel command line: BOOT_IMAGE=(hd0,msdos1)/vmlinuz-5.8.16-300.fc33.x86_64 root=UUID=b0f449d5-044c-4de0-a6de-02768c4dbd6a ro rootflags=subvol=root rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau nvidia-drm.modeset=1
    Oct 29 12:18:42 localhost.localdomain dracut-cmdline[289]: Using kernel command line parameters: rd.driver.pre=btrfs BOOT_IMAGE=(hd0,msdos1)/vmlinuz-5.8.16-300.fc33.x86_64 root=UUID=b0f449d5-044c-4de0-a6de-02768c4dbd6a ro rootflags=subvol=root rhgb quiet rd.driver.blacklist=nouveau modprobe.blacklist=nouveau nvidia-drm.modeset=1
    Oct 29 19:18:49 localhost.localdomain kernel: nvidia-nvlink: Nvlink Core is being initialized, major device number 237
    Oct 29 19:18:49 localhost.localdomain kernel: NVRM: loading NVIDIA UNIX x86_64 Kernel Module  455.28  Wed Sep 30 01:05:16 UTC 2020
    Oct 29 19:18:53 localhost.localdomain kernel: nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  455.28  Wed Sep 30 00:58:12 UTC 2020
    Oct 29 19:18:53 localhost.localdomain kernel: [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver
    Oct 29 19:18:58 localhost.localdomain kernel: NVRM: GPU at PCI:0000:01:00: GPU-2663f3be-73bf-082a-ddb6-01a6d2278a23
    Oct 29 19:18:58 localhost.localdomain kernel: NVRM: Xid (PCI:0000:01:00): 79, pid=572, GPU has fallen off the bus.
    Oct 29 19:18:58 localhost.localdomain kernel: NVRM: GPU 0000:01:00.0: GPU has fallen off the bus.
    Oct 29 19:18:58 localhost.localdomain kernel: NVRM: A GPU crash dump has been created. If possible, please run
                                                  NVRM: nvidia-bug-report.sh as root to collect this data before
                                                  NVRM: the NVIDIA kernel module is unloaded.
3

P.S your pastebin is useless as it will be deleted in 24 hours, use forum attachment instead.

4

yeah, my GPU falling off the bus sounds like its going to be an insurmountable issue, i’m seeing some google traffic about a ‘persistance mode’ but not having any luck enabling that in Fedora

5

Well I will happily eat crow, I gracefully shut down my machine, opened her up and resat the GPU, and it boots!