work on err msg, add hint for common problem

Signed-off-by: Dr. Jan-Philip Gehrcke <[email protected]>

Author: jgehrcke

hack/kubelet-plugin-prestart.sh

@@ -16,28 +16,36 @@ fi
 
 echo "NVIDIA_DRIVER_ROOT (path on host): ${NVIDIA_DRIVER_ROOT}"
 
-# Note: the following path is in-container, after chroot to /driver-root (and
-# /driver-root is where NVIDIA_DRIVER_ROOT is mounted from the host filesystem).
-# This typically outputs /usr/bin/nvidia-smi.
-echo "command -v nvidia-smi: $(command -v nvidia-smi)"
-
 while true
 do
 if ! command -v nvidia-smi &>/dev/null
 then
     printf '%b' \
-    "Command not found: 'nvidia-smi'. " \
-    "Has the NVIDIA GPU driver been set up on the host? " \
+    "Not in PATH: 'nvidia-smi'. " \
+    "Has the NVIDIA GPU driver been set up? " \
     "The GPU driver is expected to be installed under " \
-    "NVIDIA_DRIVER_ROOT in the host filesystem. " \
-    "NVIDIA_DRIVER_ROOT is currently set to: '${NVIDIA_DRIVER_ROOT}'. " \
-    "If that value appears to be unexpected: " \
+    "NVIDIA_DRIVER_ROOT ('${NVIDIA_DRIVER_ROOT}') in the host filesystem. " \
+    "If NVIDIA_DRIVER_ROOT appears to be unexpected: " \
     "review and adjust the 'nvidiaDriverRoot' Helm chart variable. " \
     "If the value is expected: review if the GPU driver has " \
     "actually been installed under NVIDIA_DRIVER_ROOT. " \
-    "If you chose the NVIDIA GPU Operator to manage the GPU driver: " \
+    "If you chose the NVIDIA GPU Operator to manage the GPU driver " \
+    "(NVIDIA_DRIVER_ROOT set to /run/nvidia/driver): "\
     "make sure that Operator is deployed and healthy.\n"
+
+    # Provide hint for a specific, common mistake.
+    if   [ "$NVIDIA_DRIVER_ROOT" == "/" ] && [ -f /run/nvidia/driver/usr/bin/nvidia-smi ]; then
+        printf '%b' \
+        "Note: /run/nvidia/driver/usr/bin/nvidia-smi exists on the host, you " \
+        "may want to re-install the DRA driver Helm chart with " \
+        "--set nvidiaDriverRoot=/run/nvidia/driver\n"
+    fi
 else
+    # Note: the following path is in-container, after chroot to /driver-root (and
+    # /driver-root is where NVIDIA_DRIVER_ROOT is mounted from the host filesystem).
+    # This typically outputs /usr/bin/nvidia-smi.
+    echo "command -v nvidia-smi: $(command -v nvidia-smi)"
+
     # This may be slow or hang, in a bad setup.
     echo "invoking nvidia-smi"
     nvidia-smi