Merge pull request #668 from javipolo/kernel_version_check

Check for driver and kernel version mismatch
containers · Jul 11, 2024 · 5f5e6d0 · 5f5e6d0
2 parents 3d1ae62 + 0ae96f7
commit 5f5e6d0
Show file tree

Hide file tree

Showing 2 changed files with 6 additions and 1 deletion.
diff --git a/training/common/Makefile.common b/training/common/Makefile.common
@@ -43,7 +43,7 @@ MODELS_CONTAINERFILE = $(OUTDIR)/Containerfile.models
 
 DRIVER_TOOLKIT_IMAGE_NAME ?= driver-toolkit
 DRIVER_TOOLKIT_IMAGE_TAG ?= latest
-DRIVER_TOOLKIT_IMAGE = ${REGISTRY}/${REGISTRY_ORG}/${DRIVER_TOOLKIT_IMAGE_NAME}:${DRIVER_TOOLKIT_IMAGE_TAG}
+DRIVER_TOOLKIT_IMAGE ?= ${REGISTRY}/${REGISTRY_ORG}/${DRIVER_TOOLKIT_IMAGE_NAME}:${DRIVER_TOOLKIT_IMAGE_TAG}
 
 ENABLE_RT ?=
 

diff --git a/training/nvidia-bootc/Containerfile b/training/nvidia-bootc/Containerfile
@@ -132,6 +132,11 @@ RUN mv /etc/selinux /etc/selinux.tmp \
         nvidia-container-toolkit \
         rsync \
         ${EXTRA_RPM_PACKAGES} \
+    && if [[ "$(rpm -qa | grep kernel-core | wc -l)" != "1" ]]; then \
+        echo "ERROR - Multiple kernel-core packages detected"; \
+        echo "This usually means that nvidia-drivers are built for a different kernel version than the one installed"; \
+        exit 1; \
+       fi \
     && if [ "$DRIVER_TYPE" != "vgpu" ] && [ "$TARGET_ARCH" != "arm64" ]; then \
         versionArray=(${DRIVER_VERSION//./ }); \
         DRIVER_BRANCH=${versionArray[0]}; \