docs: add PodGroup anti-affinity examples for inference placement across HyperNodes
Maintainers usually reply within 1 day
@SoumaCloud is already working on this.
Since Nov 5, 2026.
Assessment
This issue has not been assessed yet.
Description
Description
Part of:https://github.com/volcano-sh/website/issues/595
Add documentation demonstrating how to use PodGroup anti-affinity to distribute inference workloads across different HyperNodes in a Volcano cluster.
The guide should explain how the group-topology-affinity plugin works, how it can be configured, and how PodGroup anti-affinity can be used to prevent multiple inference instances from being placed on the same HyperNode.
Scope of Work
- PodGroup anti-affinity: Explain its purpose and how it differs from regular Pod anti-affinity.
- group-topology-affinity plugin: Document its functionality, configuration options, and role in the scheduling process.
- HyperNode-aware placement: Demonstrate how to spread multiple inference PodGroups across different HyperNodes.
- Plugin configuration: Provide the required scheduler configuration and relevant PodGroup specifications.
- Practical examples: Provide complete Kubernetes manifests for a representative inference workload.
- Scheduling behaviour: Explain how Volcano handles placement constraints and what happens when the required topology cannot be satisfied.
Validation Requirements
All examples must be tested on a running Kubernetes and Volcano cluster.
Validation must demonstrate actual scheduling behaviour, not just successful YAML parsing or resource creation.
The documentation PR should include:
- Kubernetes and Volcano versions tested.
- Required scheduler plugin configuration.
- Complete workload and PodGroup manifests.
- Commands used to deploy and verify the example.
- Observed PodGroup and Pod placement across HyperNodes.
- Relevant scheduler events or logs demonstrating the expected behaviour.
A local Kind cluster can be used if it supports the required HyperNode topology and scheduling behaviour.
Expected Outcome
A practical, implementation-backed documentation guide that enables users to configure and verify PodGroup anti-affinity for distributing inference workloads across HyperNodes.
The documentation should be added to the appropriate scheduling features section of the Volcano website, with links to relevant plugin configuration references.
Parent Issue
- Dominant language
- JavaScript
- Stars
- 33
- Forks
- 137
- Avg merge
- 3d 23h
- Merged PRs (30d)
- 4
Getting set up
- No Dockerfile or Docker Compose file
- Has a pull request template
- No contributing guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
More from volcano-sh/website
-
Typo in the load aware descheduler sectionPossibly taken A pull request linked to this issue is open or already merged. Open
Difficulty 1/5 Under an hour Newbie friendliness 78/100
volcano-sh/website#585 · 1 comment ·
Maintainers usually reply within 1 day
-
docs: audit scheduler and job plugins and add documentation indexPossibly taken @SoumaCloud claimed this 8 days ago. Open
volcano-sh/website#600 · 1 comment · 1 assignee ·
Maintainers usually reply within 1 day
-
docs: refresh homepage ecosystem cards with KubeRay, Grove, and LeaderWorkerSetPossibly taken @lmilbaum claimed this 5 days ago. Open
Difficulty 3/5 1-2 days Newbie friendliness 65/100
volcano-sh/website#597 ·
Maintainers usually reply within 1 day
-
[Umbrella] Refresh website architecture, key features, plugins, and ecosystem docsPossibly taken @JesseStutler claimed this 11 days ago. Open
volcano-sh/website#595 · 12 comments · 1 assignee ·
Maintainers usually reply within 1 day
-
docs: fix navbar Slack link/a11y, Spark Operator URLs, and colocation imagesPossibly taken @kashmeertl claimed this 94 days ago. Open
Difficulty 3/5 1-2 days Newbie friendliness 68/100
volcano-sh/website#570 ·
Maintainers usually reply within 1 day
All issues in volcano-sh/website
Similar issues
-
Difficulty 2/5 1-3 hours Newbie friendliness 79/100
Maintainers usually reply within 1 day
-
good first issue tests
Difficulty 1/5 1-3 hours Newbie friendliness 68/100
Maintainers usually reply within 1 day
-
bug role: dev-lead
Difficulty 2/5 1-3 hours Newbie friendliness 66/100
briangwalker/TrussExplorer#424 ·
Maintainers usually reply within 1 day
-
Difficulty 2/5 1-3 hours Newbie friendliness 80/100
eliasstravik/herdr-projects#1032 ·
-
enhancement good first issue
Difficulty 2/5 1-3 hours Newbie friendliness 90/100