Find a file
Aleksei Sviridkin f45facca77
feat(hami): add HAMi GPU virtualization system package (#2484)
## What this PR does

Documentation: https://github.com/cozystack/website/pull/517

Integrates [HAMi](https://github.com/Project-HAMi/HAMi) v2.8.1 (CNCF
Sandbox) into Cozystack as a system-level package for fractional GPU
sharing in tenant Kubernetes clusters.

The integration covers three layers:

- **System chart** (`packages/system/hami/`): Vendors upstream HAMi Helm
chart with device plugin, scheduler extender, mutating webhook, and
RuntimeClass configuration. The broken DRA subchart was removed — it
renders resources even when disabled and references unpublished images.

- **Kubernetes app addon** (`packages/apps/kubernetes/`): HAMi exposed
as an optional toggle (`hami.enabled`). Automatically disables GPU
Operator's native device plugin when active to avoid conflicts. Enforces
hard dependency on GPU Operator.

- **Platform registration** (`packages/core/platform/`): HAMi declared
as PackageSource with gpu-operator dependency, included in the iaas
bundle.

**Known limitation**: HAMi-core relies on a private glibc symbol
(`_dl_sym`) removed in glibc 2.34, which breaks compute isolation on
modern container images (Ubuntu 22.04+, PyTorch/TensorFlow official
images). Alpine/musl is entirely incompatible. See the package README
for details and upstream issue references.

### Screenshots

N/A — no UI changes.

### Release note

```release-note
feat(hami): add HAMi GPU virtualization as an optional system package for fractional GPU sharing. Enables memory and compute isolation for NVIDIA GPUs across tenant workloads. Integrates with GPU Operator and can be enabled per-cluster via the hami.enabled toggle. Note: compute isolation requires glibc < 2.34 in workload containers.
```

<!-- This is an auto-generated comment: release notes by coderabbit.ai
-->
## Summary by CodeRabbit

* **New Features**
* Added HAMi GPU virtualization middleware as an optional cluster addon
to enable fractional GPU sharing.
* HAMi addon includes an enable toggle (default: disabled) and a
customizable Helm values override; when enabled it enforces GPU Operator
presence and adjusts GPU Operator values accordingly.
* New HelmRelease integration for deploying HAMi with dependency
ordering and conditional rendering.

* **Documentation**
* Added HAMi setup guide, compatibility notes, and updated cluster addon
configuration reference.
<!-- end of auto-generated comment: release notes by coderabbit.ai -->
2026-04-29 16:22:58 +03:00
.gemini docs: adopt Conventional Commits across contributing docs 2026-04-13 22:50:08 +03:00
.github ci(api): add codegen drift check (#2463) 2026-04-28 12:41:48 +05:00
api chore(kubernetes): regenerate CRD after merge conflict resolution 2026-04-28 19:02:35 +03:00
cmd fix(kube-ovn): scope kubeovn-plunger cache and RBAC to its namespace 2026-04-21 17:13:27 +02:00
dashboards fix(monitoring): filter zero-valued series from active tenants count 2026-04-28 12:15:32 +03:00
docs docs(changelog): add 1.3.1 entries for backports landed after auto-gen 2026-04-29 12:46:18 +05:00
examples/backups/vmi feat(backups): restore vmi to copy in another namespace 2026-04-09 14:06:33 +04:00
hack feat(monitoring): add GPU observability dashboards and recording rules (#2418) 2026-04-28 22:42:01 +03:00
img Cozystack logo for dark GitHub theme (#9) 2024-02-09 11:02:41 +01:00
internal feat(operator): add per-package upgradeCRDs policy for HelmRelease 2026-04-27 11:19:29 +03:00
packages feat(hami): add HAMi GPU virtualization system package (#2484) 2026-04-29 16:22:58 +03:00
pkg fix(api): prevent IDOR in TenantNamespace Get and Watch handlers (#2471) 2026-04-28 19:07:07 +05:00
tools/openapi-gen [docs] Added openapi generation tool 2026-03-25 15:57:25 +05:00
.coderabbit.yaml [ci] Enable CodeRabbit incremental reviews 2026-04-13 13:32:15 +03:00
.gitignore Revert "chore: ignore CLAUDE.local.md" 2026-04-28 11:43:30 +03:00
.pre-commit-config.yaml ci(api): broaden codegen drift trigger paths and detect untracked files 2026-04-23 22:19:35 +05:00
ADOPTERS.md Update ADOPTERS.md by adding new adopter 2025-12-18 16:45:06 +03:00
AGENTS.md docs(agents): use full path .github/labels.yml in AGENTS.md 2026-04-27 03:30:43 +03:00
CODE_OF_CONDUCT.md Update CODE_OF_CONDUCT.md 2025-10-08 09:43:34 +05:00
CONTRIBUTING.md [docs] Proofread the readme and contributing 2025-04-09 11:09:19 +03:00
CONTRIBUTOR_LADDER.md Update CONTRIBUTOR_LADDER.md 2025-10-08 09:28:27 +05:00
go.mod feat(controller): add BucketClaim support to WorkloadMonitorReconciler 2026-04-17 10:58:47 +03:00
go.sum feat(controller): add BucketClaim support to WorkloadMonitorReconciler 2026-04-17 10:58:47 +03:00
GOVERNANCE.md Create GOVERNANCE.md (#733) 2025-04-01 18:48:14 +02:00
LICENSE Preapare release v0.0.1 2024-02-08 12:04:32 +01:00
MAINTAINERS.md Add Mattia Eleuteri (@mattia-eleuteri) as Maintainer 2026-04-15 23:46:43 +05:00
Makefile build(linstor): include linstor-gui in root image build target (#2498) 2026-04-28 13:55:25 +03:00
README.md Update README.md 2026-04-16 14:30:51 +05:00
SECURITY.md docs: add SECURITY.md 2026-03-17 02:57:52 +05:00

Cozystack Cozystack

Open Source Apache-2.0 License Support Active GitHub Release GitHub Commit OpenSSF Best Practices

Cozystack

Cozystack is a free platform and framework for building clouds.

Cozystack is a CNCF Sandbox Level Project that was originally built and sponsored by Ænix.

With Cozystack, you can transform a bunch of servers into an intelligent system with a simple REST API for spawning Kubernetes clusters, Database-as-a-Service, virtual machines, load balancers, HTTP caching services, and other services with ease.

Use Cozystack to build your own cloud or provide a cost-effective development environment.

Cozystack user interface

Use-Cases

Documentation

The documentation is located on the cozystack.io website.

Read the Getting Started section for a quick start.

If you encounter any difficulties, start with the troubleshooting guide and work your way through the process that we've outlined.

Versioning

Versioning adheres to the Semantic Versioning principles.
A full list of the available releases is available in the GitHub repository's Release section.

Contributions

Contributions are highly appreciated and very welcomed!

In case of bugs, please check if the issue has already been opened by checking the GitHub Issues section. If it isn't, you can open a new one. A detailed report will help us replicate it, assess it, and work on a fix.

You can express your intention to on the fix on your own. Commits are used to generate the changelog, and their author will be referenced in it.

If you have Feature Requests please use the Discussion's Feature Request section.

Community

You are welcome to join our Telegram group and come to our weekly community meetings. Add them to your Google Calendar or iCal for convenience.

License

Cozystack is licensed under Apache 2.0.
The code is provided as-is with no warranties.

Commercial Support

A list of companies providing commercial support for this project can be found on official site.