Hi,
the aim is we currently have Centos7 & want to deploy Nvidia drivers & docker containers to all machines with Ansible, but iI want to restrict install to a certain version/branch so the configuration of all are the same & upgrade process can be restricted & rolled out slowly.
this repo here is what our developers are after installing from
and what they want is to use these.
sudo yum -y install yum-plugin-nvidia-0.3
sudo yum -y install nvidia-driver-latest-dkms cuda-drivers
I believe I can do this
- yum-plugin-nvidia-(Number)
however not sure on how to do the same with dkms & cuda
Looking at the repo, there’s no “branch locking” for dkms installs, only -latest.
For branch locking, you’ll have to use the standard kernel and the pre-compiled, non-dkms driver.
???
That’s the docker container, not a driver. There are no branches.
Regarding the driver repo, that’s only a preview supporting only one older kernel.
Since there’s no further info about that, I don’t know if that ever went to production, if RH took over or this was just dropped silently. Only just noticed that you can’t even download any other driver but the .run installer directly from nvidia anymore now.
Maybe some nvidia staff member will be so kind to clear things up.
@generix sorry linked wrong repos thats my bad
the aim is we currently have Centos7 & want to deploy Nvidia drivers & docker containers to all machines with Ansible, but i want to restrict install to a certain version/branch so configuration of all are the same & upgrade process can be restrticted & rolled out slowly.
this repo here is what our developers are after installing from
and what they want is to these these.
sudo yum -y install yum-plugin-nvidia-0.3
sudo yum -y install nvidia-driver-latest-dkms cuda-drivers
i believe i can do this
- yum-plugin-nvidia-(Number)
- yum cuda-drivers-(Number)
however not sure on dkms & cuda