Skip to content

Conversation

@abuccts
Copy link
Member

@abuccts abuccts commented Oct 26, 2025

Fix service build and deployment on arm64 architecture.

Fix service build and deployment on arm64 architecture.
Copy link

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull Request Overview

This PR enables deployment on arm64 architecture by updating Docker configurations and Kubernetes deployment templates to support multi-architecture builds and private container registries.

Key changes:

  • Adds imagePullSecrets configuration to deployment templates for private registry authentication
  • Creates new Dockerfiles for device plugins and kube-scheduler to enable custom image builds
  • Updates deployment scripts to use configurable registry prefixes and image tags

Reviewed Changes

Copilot reviewed 13 out of 13 changed files in this pull request and generated 5 comments.

Show a summary per file
File Description
src/hivedscheduler/deploy/hivedscheduler.yaml.template Adds imagePullSecrets for private registry access
src/hivedscheduler/build/kube-scheduler.k8s.dockerfile New Dockerfile for kube-scheduler image
src/hivedscheduler/build/hivedscheduler.k8s.dockerfile New Dockerfile replacing the common dockerfile with updated license header
src/hivedscheduler/build/hivedscheduler.common.dockerfile Removed in favor of k8s-specific dockerfile
src/frameworkcontroller/deploy/frameworkcontroller.yaml.template Adds imagePullSecrets for private registry access
src/device-plugin/deploy/start.sh.template Updates to use configurable registry and adds imagePullSecrets injection
src/device-plugin/deploy/service.yaml Adds device-plugin.yaml to template list
src/device-plugin/deploy/device-plugin.yaml.template Replaces hardcoded image with configurable registry and adds imagePullSecrets
src/device-plugin/build/k8s-rocm-device-plugin.k8s.dockerfile New Dockerfile for AMD ROCm device plugin
src/device-plugin/build/k8s-rdma-shared-dev-plugin.k8s.dockerfile New Dockerfile for RDMA device plugin
src/device-plugin/build/k8s-nvidia-device-plugin.k8s.dockerfile New Dockerfile for NVIDIA device plugin
src/device-plugin/build/k8s-host-device-plugin.k8s.dockerfile New Dockerfile for host device plugin
contrib/kubespray/script/environment.sh Adds force flag to symlink commands to prevent errors on re-runs

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

Update.
Use buildkit for Docker build.
Upgrade NVIDIA k8s device plugin version.
@yukirora
Copy link

yukirora commented Nov 4, 2025

@zhogu could you please help check and fix the CI error in this PR?

@yukirora yukirora mentioned this pull request Nov 4, 2025
26 tasks
Update docker install.
@abuccts abuccts requested a review from a team as a code owner November 7, 2025 22:41
Add ghcr login.
Update.
@abuccts abuccts force-pushed the xiongyf/fix-deploy-arm branch from 5c5cd42 to a93459b Compare November 8, 2025 00:03
@abuccts abuccts merged commit 7e1a263 into dev Nov 8, 2025
3 checks passed
@abuccts abuccts deleted the xiongyf/fix-deploy-arm branch November 8, 2025 00:35
@yukirora yukirora mentioned this pull request Nov 11, 2025
25 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

4 participants