NVIDIA / libnvidia-container

NVIDIA container runtime library
Apache License 2.0
843 stars 205 forks source link

RHEL 8.8 support #208

Open seeteena opened 1 year ago

seeteena commented 1 year ago

Hello,

I am attempting to install libnvidia-container on RHEL 8.8 - any idea if/when there will be a package built for it in the repo?

distribution=$(. /etc/os-release;echo $ID$VERSION_ID) && curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.repo | sudo tee /etc/yum.repos.d/nvidia-container-toolkit.repo
Unsupported distribution!
Check https://nvidia.github.io/libnvidia-container

Thanks!

I can use RHEL 8.7 not RHEL 8.8

[root@ltciofvtr-spoon4 ~]# distribution=rhel8.7 [root@ltciofvtr-spoon4 ~]# curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.repo | sudo tee /etc/yum.repos.d/libnvidia-container.repo [libnvidia-container] name=libnvidia-container baseurl=https://nvidia.github.io/libnvidia-container/stable/centos8/$basearch repo_gpgcheck=1 gpgcheck=0 enabled=1 gpgkey=https://nvidia.github.io/libnvidia-container/gpgkey sslverify=1 sslcacert=/etc/pki/tls/certs/ca-bundle.crt

[libnvidia-container-experimental] name=libnvidia-container-experimental baseurl=https://nvidia.github.io/libnvidia-container/experimental/centos8/$basearch repo_gpgcheck=1 gpgcheck=0 enabled=0 gpgkey=https://nvidia.github.io/libnvidia-container/gpgkey sslverify=1 sslcacert=/etc/pki/tls/certs/ca-bundle.crt [root@ltciofvtr-spoon4 ~]# distribution=rhel8.8 [root@ltciofvtr-spoon4 ~]# curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.repo | sudo tee /etc/yum.repos.d/libnvidia-container.repo

Unsupported distribution!

Check https://nvidia.github.io/libnvidia-container

[root@ltciofvtr-spoon4 ~]#

seeteena commented 1 year ago

Any update

klueska commented 1 year ago

We don’t build packages for each system (because the distributed binaries’ only dependency is glibc). All RedHat packages are just a symlink to the centos8 packages.

So if/when we add support for RHEL 8.8 it would just mean that we added it to our test matrix (but would still be using the centos8 packages for it).

Meaning, feel free to use the RHEL 8.7 package as nothing will likely change except our test coverage.

elezar commented 1 year ago

@seeteena we are in the last steps of distributing our package though the CUDA download repositories for ubuntu and rhel8-based distributions. One this happens it should not be required to configure additional repos for any of the components of the NVIDIA Container Toolkit.

Pavithra1602 commented 1 year ago

getting below error If i try to pull container images, please let me know if i am missing any steps.

[root@ltciofvtr-spoon4 ~]# podman pull docker.io/nvidia/cuda-ppc64le Trying to pull docker.io/nvidia/cuda-ppc64le:latest... Error: initializing source docker://nvidia/cuda-ppc64le:latest: reading manifest latest in docker.io/nvidia/cuda-ppc64le: manifest unknown

seeteena commented 1 year ago

Any update here?

elezar commented 1 year ago

@seeteena were you able to install the centos8 or rhel8.7 packages?

distribution=centos8 && curl -s -L https://nvidia.github.io/libnvidia-container/$distribution/libnvidia-container.repo | sudo tee /etc/yum.repos.d/nvidia-container-toolkit.repo

Note that the v1.13.1 of the NVIDIA Contianer Toolkit and its dependencies should now be available on the CUDA downloads repositories, meaning that if you have these configured they can be installed directly.

Pavithra1602 commented 1 year ago

With above repo i am able to install nvidia-container-toolkit.

m.repos.d/nvidia-container-toolkit.repobnvidia-container.repo | sudo tee /etc/yum [libnvidia-container] name=libnvidia-container baseurl=https://nvidia.github.io/libnvidia-container/stable/centos8/$basearch repo_gpgcheck=1 gpgcheck=0 enabled=1 gpgkey=https://nvidia.github.io/libnvidia-container/gpgkey sslverify=1 sslcacert=/etc/pki/tls/certs/ca-bundle.crt

[libnvidia-container-experimental] name=libnvidia-container-experimental baseurl=https://nvidia.github.io/libnvidia-container/experimental/centos8/$basearch repo_gpgcheck=1 gpgcheck=0 enabled=0 gpgkey=https://nvidia.github.io/libnvidia-container/gpgkey sslverify=1 sslcacert=/etc/pki/tls/certs/ca-bundle.crt [root@ltc-wspoon12 ~]# -container-toolkit ~]# sudo dnf clean expire-cache && sudo dnf install -y nvidia- Updating Subscription Management repositories. Unable to read consumer identity

This system is not registered with an entitlement server. You can use subscription-manager to register.

Cache was expired 0 files removed Updating Subscription Management repositories. Unable to read consumer identity

This system is not registered with an entitlement server. You can use subscription-manager to register.

libnvidia-container 2.5 kB/s | 833 B 00:00
libnvidia-container 45 kB/s | 3.1 kB 00:00
Importing GPG key 0xF796ECB0: Userid : "NVIDIA CORPORATION (Open Source Projects) cudatools@nvidia.com" Fingerprint: C95B 321B 61E8 8C18 09C4 F759 DDCA E044 F796 ECB0 From : https://nvidia.github.io/libnvidia-container/gpgkey libnvidia-container 120 kB/s | 44 kB 00:00
rh82_base_upd 3.2 kB/s | 2.8 kB 00:00
rh82_base_upd 778 kB/s | 2.0 MB 00:02
rh82_app_upd 3.2 kB/s | 3.2 kB 00:01
rh82_app_upd 1.8 MB/s | 7.1 MB 00:03
rh82_app_upd 3.5 kB/s | 3.2 kB 00:00
rh82_app_upd 694 kB/s | 1.8 MB 00:02
Dependencies resolved.

Package Arch Version Repository Size

Installing: nvidia-container-toolkit ppc64le 1.13.2-1 libnvidia-container 791 k Installing dependencies: libnvidia-container-tools ppc64le 1.13.2-1 libnvidia-container 57 k libnvidia-container1 ppc64le 1.13.2-1 libnvidia-container 920 k nvidia-container-toolkit-base ppc64le 1.13.2-1 libnvidia-container 2.7 M

Transaction Summary

Install 4 Packages

Total download size: 4.4 M Installed size: 15 M Downloading Packages: (1/4): libnvidia-container-tools-1.13.2-1.ppc64 396 kB/s | 57 kB 00:00
(2/4): nvidia-container-toolkit-1.13.2-1.ppc64l 3.2 MB/s | 791 kB 00:00
(3/4): libnvidia-container1-1.13.2-1.ppc64le.rp 2.2 MB/s | 920 kB 00:00
(4/4): nvidia-container-toolkit-base-1.13.2-1.p 4.3 MB/s | 2.7 MB 00:00

Total 5.4 MB/s | 4.4 MB 00:00
Running transaction check Transaction check succeeded. Running transaction test Transaction test succeeded. Running transaction Preparing : 1/1 Installing : nvidia-container-toolkit-base-1.13.2-1.ppc64le 1/4 Installing : libnvidia-container1-1.13.2-1.ppc64le 2/4 Running scriptlet: libnvidia-container1-1.13.2-1.ppc64le 2/4 /sbin/ldconfig: /usr/lib64/llvm15/lib/libclang.so.15 is not a symbolic link

Installing : libnvidia-container-tools-1.13.2-1.ppc64le 3/4 Installing : nvidia-container-toolkit-1.13.2-1.ppc64le 4/4 Running scriptlet: nvidia-container-toolkit-1.13.2-1.ppc64le 4/4 /sbin/ldconfig: /usr/lib64/llvm15/lib/libclang.so.15 is not a symbolic link

Verifying : libnvidia-container-tools-1.13.2-1.ppc64le 1/4 Verifying : libnvidia-container1-1.13.2-1.ppc64le 2/4 Verifying : nvidia-container-toolkit-1.13.2-1.ppc64le 3/4 Verifying : nvidia-container-toolkit-base-1.13.2-1.ppc64le 4/4 Installed products updated.

Installed: libnvidia-container-tools-1.13.2-1.ppc64le
libnvidia-container1-1.13.2-1.ppc64le
nvidia-container-toolkit-1.13.2-1.ppc64le
nvidia-container-toolkit-base-1.13.2-1.ppc64le

Complete!

Could pull other container images like busybox, nut not ale to pull nvidia related containers [ Ex: cuda-ppc64le ]

[root@ltc-wspoon12 ~]# podman pull docker.io/ppc64le/busybox Trying to pull docker.io/ppc64le/busybox:latest... Getting image source signatures Writing manifest to image destination Storing signatures 0ff8a9aafa0205be76e4ea440ddf3a69295e393cc39b3eea289c642787d5611b

[root@ltc-wspoon12 ~]# podman pull docker.io/nvidia/cuda-ppc64le Trying to pull docker.io/nvidia/cuda-ppc64le:latest... Error: initializing source docker://nvidia/cuda-ppc64le:latest: reading manifest latest in docker.io/nvidia/cuda-ppc64le: manifest unknown

Thanks, Pavithra

seeteena commented 1 year ago

any update here?

SACHIN-BAPPALIGE commented 12 months ago

RHEL 8.9 nvidia package not available . Please let's know the package details

elezar commented 12 months ago

With the NVIDIA Container Toolkit v1.14.0 release we revamped how are packages are built and distributed.

Please follow the updated instructions here. These should also work for RHEL 8.9.

SACHIN-BAPPALIGE commented 12 months ago

Could you please check this error

Please follow the updated instructions here. These should also work for RHEL 8.9.


cat /etc/redhat-release
Red Hat Enterprise Linux release 8.9 Beta (Ootpa)

curl -s -L https://nvidia.github.io/libnvidia-container/stable/rpm/nvidia-container-toolkit.repo | \

sudo tee /etc/yum.repos.d/nvidia-container-toolkit.repo [nvidia-container-toolkit] name=nvidia-container-toolkit baseurl=https://nvidia.github.io/libnvidia-container/stable/rpm/$basearch repo_gpgcheck=1 gpgcheck=0 enabled=1 gpgkey=https://nvidia.github.io/libnvidia-container/gpgkey sslverify=1 sslcacert=/etc/pki/tls/certs/ca-bundle.crt

[nvidia-container-toolkit-experimental] name=nvidia-container-toolkit-experimental baseurl=https://nvidia.github.io/libnvidia-container/experimental/rpm/$basearch repo_gpgcheck=1 gpgcheck=0 enabled=0 gpgkey=https://nvidia.github.io/libnvidia-container/gpgkey sslverify=1 sslcacert=/etc/pki/tls/certs/ca-bundle.crt

ls

nvidia-container-toolkit.repo redhat.repo rhel.repo

cat nvidia-container-toolkit.repo

[nvidia-container-toolkit] name=nvidia-container-toolkit baseurl=https://nvidia.github.io/libnvidia-container/stable/rpm/$basearch repo_gpgcheck=1 gpgcheck=0 enabled=1 gpgkey=https://nvidia.github.io/libnvidia-container/gpgkey sslverify=1 sslcacert=/etc/pki/tls/certs/ca-bundle.crt

[nvidia-container-toolkit-experimental] name=nvidia-container-toolkit-experimental baseurl=https://nvidia.github.io/libnvidia-container/experimental/rpm/$basearch repo_gpgcheck=1 gpgcheck=0 enabled=0 gpgkey=https://nvidia.github.io/libnvidia-container/gpgkey sslverify=1 sslcacert=/etc/pki/tls/certs/ca-bundle.crt

dnf install -y nvidia-container-toolkit

Updating Subscription Management repositories. Unable to read consumer identity This system is not registered with an entitlement server. You can use subscription-manager to register. nvidia-container-toolkit 30 kB/s | 4.8 kB 00:00 Errors during downloading metadata for repository 'nvidia-container-toolkit':

elezar commented 12 months ago

@SACHIN-BAPPALIGE I see that you're using ppc64le. Those packages are not yet in the common repo.

For ppc64le please use:

https://nvidia.github.io/libnvidia-container/stable/centos8

instead of

https://nvidia.github.io/libnvidia-container/stable/rpm
SACHIN-BAPPALIGE commented 8 months ago

@elezar That worked . Thanks