[VL][CI] Update gpu image to centos-9-jdk17-cuda13.1-cudf

[marin-ma]

The jdk version in the gpu image is now jdk17 after #11835

[Copilot]

Pull request overview

Warning

Copilot couldn't run its full agentic review because it didn't start before the timeout. Make sure your repository has a runner available, or add a copilot-code-review.yml file specifying one with the runs-on attribute. See the docs for more details.

Updates the Velox GPU (cuDF) Docker image references and CI workflow configuration to use the CentOS 9 + JDK 17 tag.

Changes:

• Replaced centos-9-jdk8-cudf image tag references with centos-9-jdk17-cudf across docs, scripts, and CI workflows.
• Renamed the Docker image build workflow job and adjusted the pushed tag to publish the JDK17 variant.
• Extended the cuDF Dockerfile build to enable additional cloud filesystem features (S3/GCS/ABFS).

Reviewed changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
docs/get-started/VeloxGPU.md	Updates GPU image pull/run examples to the JDK17 cuDF image tag.
docs/developers/velox-backend-CI.md	Updates documented GPU image tag to JDK17 and fixes formatting indentation.
dev/start-cudf.sh	Updates the example container run command to use the JDK17 cuDF image tag.
dev/docker/cudf/Dockerfile	Switches the bundle build to enable additional cloud filesystem integrations alongside GPU.
.github/workflows/velox_backend_x86.yml	Updates CI build container image to centos-9-jdk17-cudf.
.github/workflows/velox_backend_cache.yml	Updates cache build container image to centos-9-jdk17-cudf.
.github/workflows/docker_image.yml	Renames the cuDF build job and updates the pushed Docker tag to JDK17.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 7 out of 7 changed files in this pull request and generated no new comments.

Suppressed comments (1)

dev/docker/cudf/Dockerfile:31

• This PR is described as updating the cuDF GPU image from JDK 8 to JDK 17, but this change also enables S3/GCS/ABFS in the Docker image build. That’s a functional expansion that can change the produced artifact (size, deps, build time, included features). If intentional, please update the PR title/description (or add a short note in the relevant docs/CI doc) to reflect the new backends; if not, consider reverting these flags and moving them to a dedicated PR.

    bash ./dev/buildbundle-veloxbe.sh --run_setup_script=ON --build_arrow=ON --spark_version=3.5 --build_tests=ON --build_benchmarks=ON --enable_s3=ON --enable_gcs=ON --enable_abfs=ON --enable_gpu=ON && \

[Copilot]

Pull request overview

Copilot reviewed 5 out of 5 changed files in this pull request and generated no new comments.

Suppressed comments (3)

docs/get-started/VeloxGPU.md:42

• The PR title/description says this is a JDK update, but the new tag also introduces an explicit CUDA version (cuda13.1). Please either (a) update the PR title/description to mention the CUDA version bump as part of the change, or (b) switch to an image tag that reflects only the intended JDK change (if one exists). This avoids under-communicating a potentially breaking runtime requirement change.

docker pull apache/gluten:centos-9-jdk17-cuda13.1-cudf  # Pre-built GPU image
docker run --name gpu_gluten_container --gpus all -it apache/gluten:centos-9-jdk17-cuda13.1-cudf

docs/get-started/VeloxGPU.md:42

• Since the image tag now encodes a specific CUDA version (cuda13.1), the doc should mention expected NVIDIA driver/CUDA compatibility requirements (and/or link to where those requirements are defined). Without that, users may follow the instructions and hit runtime failures due to incompatible host drivers.

docker pull apache/gluten:centos-9-jdk17-cuda13.1-cudf  # Pre-built GPU image
docker run --name gpu_gluten_container --gpus all -it apache/gluten:centos-9-jdk17-cuda13.1-cudf

.github/workflows/velox_backend_x86.yml:1155

• The GPU image tag is now duplicated across multiple workflows/docs/scripts. In the GitHub workflows, consider defining a single env var (e.g., GLUTEN_CUDF_IMAGE) at workflow/job level and referencing it from each docker run. That reduces churn and the risk of partial updates if the tag changes again.

          docker run -v $GITHUB_WORKSPACE:/work -w /work apache/gluten:centos-9-jdk17-cuda13.1-cudf bash -c "