[Fix] Add CUDA-enabled Spark image

[gtopper]

📝 Description

Adds the CUDA-enabled Spark image required when MLRun requests GPU resources for Spark drivers or executors.

The image matches the existing CE Spark 3.5.6 stack while adding CUDA 12.8.1 and cuDNN 9.8.

---

🛠️ Changes Made

• Added docker/spark/Dockerfile.cuda based on nvidia/cuda:12.8.1-cudnn-devel-ubuntu22.04.
• Shared Python 3.11, connector JAR, and filesystem setup through docker/spark/scripts/ce-customize.sh.
• Added CUDA build and validation targets to docker/spark/Makefile.
• Added local checks for Spark, Scala, Java, Python, connector JARs, CUDA, cuDNN, and NVIDIA container metadata.
• Documented image build, validation, and manual GCR publication in docker/spark/README.md.
• Bumped the chart version from 0.12.0-rc.10 to 0.12.0-rc.11.

---

✅ Checklist

• I have tested the changes in this PR
• I confirmed whether my changes require a change in documentation and if so, I created another PR in MLRun for the relevant documentation.
• I confirmed whether my changes require a changes in QA tests, for example: credentials changes, resources naming change and if so, I updated the relevant Jira ticket for QA.
• I increased the Chart version in charts/mlrun-ce/Chart.yaml.
• I confirmed that the installation works both on a local Docker Desktop environment and on a real cluster when using the required prerequisites.
  • If installation issues were found, I updated the relevant Jira ticket with the issue and steps to reproduce, or updated the prerequisites documentation if the issue is related to missing or outdated prerequisites.
• If needed, update https://github.com/mlrun/ce/blob/development/charts/mlrun-ce/README.md with the relevant installation instructions and version Matrix.
• If needed, update the following values files for multi namespace support:
  • Admin values
  • User values Node Port
  • User values ClusterIP

---

🧪 Testing

• Built the linux/amd64 CUDA image with make -C docker/spark build-cuda.
• Ran make -C docker/spark validate-cuda.
• Verified Spark 3.5.6, Scala 2.12, Java 17, Python 3.11, connector JARs, CUDA 12.8.1, cuDNN 9.8, the Spark entrypoint, and NVIDIA container metadata.
• GPU device visibility and Spark driver/executor startup still require validation in an ORIS GPU environment.

---

🔗 References

• Ticket link: ORIS-4194
• External links: Apache Spark Docker images, NVIDIA CUDA container images
• Design docs links (Optional): N/A

---

🚨 Breaking Changes?

• Yes (explain below)
• No

---

🔍️ Additional Notes

• The CUDA image is published manually because this repository has no provisioned GCR publishing credentials.
• Adding the image to the Mlefi release manifest and validating the GPU Spark scenario are follow-up tasks.