From c38c6a5dc15489b889602c39686ff8cc5d01b967 Mon Sep 17 00:00:00 2001 From: Abigail McCarthy <20771501+a-mccarthy@users.noreply.github.com> Date: Wed, 27 May 2026 10:10:26 -0400 Subject: [PATCH 1/3] Update cdi notes Signed-off-by: Abigail McCarthy <20771501+a-mccarthy@users.noreply.github.com> --- gpu-operator/cdi.rst | 5 +++++ gpu-operator/getting-started.rst | 11 +++++++++-- 2 files changed, 14 insertions(+), 2 deletions(-) diff --git a/gpu-operator/cdi.rst b/gpu-operator/cdi.rst index 468029eea..913a12d2c 100644 --- a/gpu-operator/cdi.rst +++ b/gpu-operator/cdi.rst @@ -153,6 +153,11 @@ disable CDI and use the legacy NVIDIA Container Toolkit stack instead with the f About the Node Resource Interface (NRI) Plugin ********************************************** +.. note:: + + The containerd project has not yet released a general availability (GA) version of the NRI Plugin. The implementation might change before the GA release. + Refer to the `containerd NRI repository `_ for details on project details. + Node Resource Interface (NRI) is a standardized interface for plugging in extensions, called NRI Plugins, to OCI-compatible container runtimes like containerd. NRI Plugins serve as hooks which intercept pod and container lifecycle events and perform functions including injecting devices to a container, topology aware placement strategies, and more. For more details on NRI, refer to the `NRI overview `_ in the containerd repository. diff --git a/gpu-operator/getting-started.rst b/gpu-operator/getting-started.rst index 8cf8776f1..46644d006 100644 --- a/gpu-operator/getting-started.rst +++ b/gpu-operator/getting-started.rst @@ -157,6 +157,8 @@ To view all the options, run ``helm show values nvidia/gpu-operator``. * - ``cdi.nriPluginEnabled`` - When set to ``true``, the Node Resource Interface (NRI) Plugin will be used for injecting GPUs into workload containers. + The NRI Plugin is not GA and not recommended for production use. + In NRI Plugin mode, the NVIDIA Container Toolkit will no longer modify the runtime config. This feature requires containerd v1.7.30, v2.1.x, or v2.2.x. Refer to the :doc:`cdi` page for more information. @@ -584,8 +586,13 @@ For Rancher Kubernetes Engine 2 (RKE2), refer to `Deploy NVIDIA Operator `__ in the RKE2 documentation. -It's recommended that you enable CDI (default) and the NRI Plugin on RKE. -With both features enabled, you do not need to set ``runtimeClassName: nvidia`` in your pod spec. +The NRI Plugin is available for use on RKE2. With CDI (the default) and the NRI Plugin both enabled, you do not need to set ``runtimeClassName: nvidia`` in your pod specification, and you do not need to configure the ``CONTAINERD_CONFIG``, ``CONTAINERD_SOCKET``, or ``RUNTIME_CONFIG_SOURCE`` environment variables. + +.. note:: + + The containerd project has not yet released a general availability (GA) version of the NRI Plugin. The implementation might change before the GA release. + The containerd project has not yet released a general availability (GA) version of the NRI Plugin. The implementation might change before the GA release. + Refer to the `containerd NRI repository `_ for details on project details. Refer to the :ref:`v24.9.0-known-limitations`. From b21cb697db87207584051526c433b225268650d9 Mon Sep 17 00:00:00 2001 From: Abigail McCarthy <20771501+a-mccarthy@users.noreply.github.com> Date: Wed, 27 May 2026 10:20:13 -0400 Subject: [PATCH 2/3] Update spacing Signed-off-by: Abigail McCarthy <20771501+a-mccarthy@users.noreply.github.com> --- gpu-operator/getting-started.rst | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/gpu-operator/getting-started.rst b/gpu-operator/getting-started.rst index 46644d006..810fb0b11 100644 --- a/gpu-operator/getting-started.rst +++ b/gpu-operator/getting-started.rst @@ -156,10 +156,9 @@ To view all the options, run ``helm show values nvidia/gpu-operator``. - ``true`` * - ``cdi.nriPluginEnabled`` - - When set to ``true``, the Node Resource Interface (NRI) Plugin will be used for injecting GPUs into workload containers. - The NRI Plugin is not GA and not recommended for production use. + - When set to ``true``, the Node Resource Interface (NRI) Plugin will be used for injecting GPUs into workload containers. - In NRI Plugin mode, the NVIDIA Container Toolkit will no longer modify the runtime config. + In NRI Plugin mode, the NVIDIA Container Toolkit will no longer modify the runtime config. This feature requires containerd v1.7.30, v2.1.x, or v2.2.x. Refer to the :doc:`cdi` page for more information. - ``false`` From 7315935994613a81013665dd1bae799a724180d7 Mon Sep 17 00:00:00 2001 From: Abigail McCarthy <20771501+a-mccarthy@users.noreply.github.com> Date: Wed, 27 May 2026 11:47:32 -0400 Subject: [PATCH 3/3] Add nri details to containerd section Signed-off-by: Abigail McCarthy <20771501+a-mccarthy@users.noreply.github.com> --- gpu-operator/getting-started.rst | 13 ++++--------- 1 file changed, 4 insertions(+), 9 deletions(-) diff --git a/gpu-operator/getting-started.rst b/gpu-operator/getting-started.rst index 810fb0b11..6ecf7c6a2 100644 --- a/gpu-operator/getting-started.rst +++ b/gpu-operator/getting-started.rst @@ -513,8 +513,9 @@ Specifying Configuration Options for containerd .. note:: - It's recommended that you enable the NRI Plugin to configure the container runtime by setting ``cdi.nriPluginEnabled=true``. - When enabled, you do not need to specify the ``toolkit.env`` options and injecting GPUs into workload containers is handled by the NRI Plugin. + When you enable the NRI Plugin, you do not need to specify the ``toolkit.env`` options and injecting GPUs into workload containers is handled by the NRI Plugin. + You can enable the NRI Plugin to configure the container runtime by setting ``cdi.nriPluginEnabled=true``. + The NRI Plugin is available for use on RKE2. Refer to the :ref:`NRI Plugin ` documentation, for more information. When you use containerd as the container runtime, the following configuration @@ -585,13 +586,7 @@ For Rancher Kubernetes Engine 2 (RKE2), refer to `Deploy NVIDIA Operator `__ in the RKE2 documentation. -The NRI Plugin is available for use on RKE2. With CDI (the default) and the NRI Plugin both enabled, you do not need to set ``runtimeClassName: nvidia`` in your pod specification, and you do not need to configure the ``CONTAINERD_CONFIG``, ``CONTAINERD_SOCKET``, or ``RUNTIME_CONFIG_SOURCE`` environment variables. - -.. note:: - - The containerd project has not yet released a general availability (GA) version of the NRI Plugin. The implementation might change before the GA release. - The containerd project has not yet released a general availability (GA) version of the NRI Plugin. The implementation might change before the GA release. - Refer to the `containerd NRI repository `_ for details on project details. +With CDI (the default) and the NRI Plugin both enabled, you do not need to set ``runtimeClassName: nvidia`` in your pod specification, and you do not need to configure the ``CONTAINERD_CONFIG``, ``CONTAINERD_SOCKET``, or ``RUNTIME_CONFIG_SOURCE`` environment variables. Refer to the :ref:`v24.9.0-known-limitations`.