From a796f6d24c96c091d0d90427e28a85bda91af8b2 Mon Sep 17 00:00:00 2001 From: Doug Holt Date: Mon, 19 Oct 2026 15:55:38 -0600 Subject: [PATCH] docs(runai): point to gpu_operator_chart_version instead of hard-coding the default Signed-off-by: Doug Holt --- docs/k8s-cluster/runai.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/k8s-cluster/runai.md b/docs/k8s-cluster/runai.md index 5e1f3907a..22ba08aab 100644 --- a/docs/k8s-cluster/runai.md +++ b/docs/k8s-cluster/runai.md @@ -118,7 +118,7 @@ scheduler stays the default. caps each pod's GPU memory. It is opt-in because it needs a GPU stack the DeepOps defaults do not use: -- GPU Operator **v26.7.1 or newer** (DeepOps pins v26.7.0 by default), and +- GPU Operator **v26.7.1 or newer** (check `gpu_operator_chart_version` in `roles/nvidia-gpu-operator/defaults/main.yml` for the DeepOps default), and - an NVIDIA driver **r615 or newer** on every GPU node. The GPU Operator does not choose r615 by default; set `gpu_operator_driver_version` to an r615 release and `gpu_operator_preinstalled_nvidia_software: false` (or install an r615 host