Is this page helpful?

Post-installation Actions#

The post-installation actions must be manually performed.

Persistence Daemon#

NVIDIA is providing a user-space daemon on Linux to support persistence of driver state across CUDA job runs. The daemon approach provides a more elegant and robust solution to this problem than persistence mode. For more details on the NVIDIA Persistence Daemon, see the documentation here.

All the distribution packages contain a systemd preset to enable it automatically if installed as a dependency of other driver components. It can be restarted manually by running:

# systemctl restart nvidia-persistenced

Verify the Driver Version#

If you installed the driver, verify that the correct version of it is loaded. When the driver is loaded, the driver version can be found by executing the following command:

$ cat /proc/driver/nvidia/version

Reboot Notification#

The kernel modules and the user space components of the driver must always be of the same version. After a driver installation or upgrade, the system must be rebooted to load the new kernel modules. On distributions that use DNF, the driver packages register themselves with DNF as packages whose update requires a reboot, the same way as the kernel or systemd. The needs-restarting command of DNF then reports when a reboot is required.

The command exits with code 1 when a reboot is required and with code 0 otherwise, so it can also be used in scripts and automation tools.

DNF 4#

This applies to the following distributions:

  • Red Hat Enterprise Linux 8 / AlmaLinux 8 / Rocky Linux 8 / Oracle Linux 8

  • Red Hat Enterprise Linux 9 / AlmaLinux 9 / Rocky Linux 9 / Oracle Linux 9

  • Red Hat Enterprise Linux 10 / AlmaLinux 10 / Rocky Linux 10 / Oracle Linux 10

  • Amazon Linux 2023

The nvidia-driver and nvidia-driver-cuda packages install the files nvidia-driver.conf and nvidia-driver-cuda.conf in the /etc/dnf/plugins/needs-restarting.d/ directory. Each file contains the name of the package that suggests a reboot.

The needs-restarting command is provided by the dnf-plugins-core package:

# dnf install dnf-plugins-core

To check if a reboot is required, use the -r option:

# dnf needs-restarting -r
Core libraries or services have been updated since boot-up:
  * nvidia-driver-cuda
Reboot is required to fully utilize these updates.
More information: https://access.redhat.com/solutions/27943

Additional packages can be added by creating .conf files in the same directory, with one package name per line.

See also

dnf-needs-restarting(8) manual page.

DNF 5#

This applies to the following distributions:

  • Fedora 44

  • Azure Linux 4

The nvidia-driver and nvidia-driver-cuda packages install the files nvidia-driver.conf and nvidia-driver-cuda.conf in the /usr/share/dnf5/suggest-reboot.d/ directory. Each file contains the name of the package that suggests a reboot. Azure Linux provides only the compute part of the driver, so only nvidia-driver-cuda.conf is installed.

Note

The suggest-reboot.d directories are read by DNF 5.4.6 or later. Azure Linux 4 currently ships an earlier DNF 5 release; until it is updated, the driver packages are not reported. Check the installed version with rpm -q dnf5.

The needs-restarting command is provided by the dnf5-plugins package:

# dnf install dnf5-plugins

With DNF 5, the command checks if a reboot is required without any additional option. The -r option of DNF 4 is still accepted, but has no effect:

# dnf needs-restarting
Core libraries or services have been updated since boot-up:
  * nvidia-driver-cuda
Reboot is required to fully utilize these updates.
More information: https://access.redhat.com/solutions/27943

Additional packages can be added by creating .conf files, with one package name per line, in the /etc/dnf/suggest-reboot.d/ directory. A file in this directory replaces a file with the same name in /usr/share/dnf5/suggest-reboot.d/. The /etc/dnf/plugins/needs-restarting.d/ directory of DNF 4 is not used by DNF 5.

The same information is also used by dnf5 automatic. To reboot the system automatically after the driver has been upgraded, set the following in the [commands] section of /etc/dnf/automatic.conf:

reboot = when-needed

See also

dnf5-needs-restarting(8) and dnf5-automatic(8) manual pages.

Local Repository Removal#

Removal of the local repository installer is recommended after installation of the driver.

Amazon Linux 2023, Azure Linux 4, Fedora 44, KylinOS 11, Red Hat Enterprise Linux 8/9/10, AlmaLinux 8/9/10, Rocky Linux 8/9/10, Oracle Linux 8/9/10

# dnf remove nvidia-driver-local-repo\*

Azure Linux 3

# tdnf remove nvidia-driver-local-repo-\*

Ubuntu 22.04/24.04/24.04, Debian 12

# apt remove --purge nvidia-driver-local-repo\*

SUSE Linux Enterprise Server 15/16, OpenSUSE Leap 15/16

# zypper remove nvidia-driver-local-repo\*