Artificial Intelligence Computing Leadership from NVIDIA
ERROR: Unable to load the kernel module 'nvidia.ko'
On a new Amazon EC2 g2.2xlarge GPU instance running Ubuntu Server 14.04 LTS, I cannot [url=http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/using_cluster_computing.html#install-nvidia-driver]install the latest NVIDIA GRID driver[/url]. When trying to install the driver, I encounter the following error: [quote] ERROR: Unable to load the kernel module 'nvidia.ko'. This happens most frequently when this kernel module was built against the wrong or improperly configured kernel sources, with a version of gcc that differs from the one used to build the target kernel, or if a driver such as rivafb, nvidiafb, or nouveau is present and prevents the NVIDIA kernel module from obtaining ownership of the NVIDIA graphics device(s), or no NVIDIA GPU installed in this system is supported by this NVIDIA Linux graphics driver release. Please see the log entries 'Kernel module load error' and 'Kernel messages' at the end of the file '/var/log/nvidia-installer.log' for more information.[/quote] How can I resolve this error and proceed with installing the driver? The output below shows the OS version on the EC2 instance. [code]$ lsb_release -a No LSB modules are available. Distributor ID: Ubuntu Description: Ubuntu 14.04.2 LTS Release: 14.04 Codename: trusty[/code]
On a new Amazon EC2 g2.2xlarge GPU instance running Ubuntu Server 14.04 LTS, I cannot install the latest NVIDIA GRID driver. When trying to install the driver, I encounter the following error:

ERROR: Unable to load the kernel module 'nvidia.ko'. This happens most
frequently when this kernel module was built against the wrong or
improperly configured kernel sources, with a version of gcc that
differs from the one used to build the target kernel, or if a driver
such as rivafb, nvidiafb, or nouveau is present and prevents the
NVIDIA kernel module from obtaining ownership of the NVIDIA graphics
device(s), or no NVIDIA GPU installed in this system is supported by
this NVIDIA Linux graphics driver release.

Please see the log entries 'Kernel module load error' and 'Kernel
messages' at the end of the file '/var/log/nvidia-installer.log' for
more information.


How can I resolve this error and proceed with installing the driver? The output below shows the OS version on the EC2 instance.

$ lsb_release -a
No LSB modules are available.
Distributor ID: Ubuntu
Description: Ubuntu 14.04.2 LTS
Release: 14.04
Codename: trusty

#1
Posted 01/17/2016 04:30 AM   
Here's the [url=http://pastie.org/private/frsz882ei85vphykvtfvg]nvidia-installer.log[/url] mentioned in the error message.
Here's the nvidia-installer.log mentioned in the error message.

#2
Posted 01/17/2016 04:41 AM   
Can't read the .log the connection times out. Have you disabled the nouveau driver? have you checked the gcc version used by the driver package (installed on the Server) is the same one used to create the server kernel? You may need to change the version of gcc in the VM to match. You may also find this useful http://tleyden.github.io/blog/2014/10/25/cuda-6-dot-5-on-aws-gpu-instance-running-ubuntu-14-dot-04/
Can't read the .log the connection times out.

Have you disabled the nouveau driver?

have you checked the gcc version used by the driver package (installed on the Server) is the same one used to create the server kernel?

You may need to change the version of gcc in the VM to match.

You may also find this useful

http://tleyden.github.io/blog/2014/10/25/cuda-6-dot-5-on-aws-gpu-instance-running-ubuntu-14-dot-04/

Jason Southern, Regional Lead for ProVis Sales - EMEA: NVIDIA Ltd.

#3
Posted 01/18/2016 10:06 PM   
Scroll To Top

Add Reply