[hami][WIP] Draft HAMI project integration#1844
[hami][WIP] Draft HAMI project integration#1844tabu-a wants to merge 2 commits intocozystack:maincozystack/cozystack:mainfrom
Conversation
|
Important Review skippedDraft detected. Please check the settings in the CodeRabbit UI or the You can disable this status message by setting the Note Other AI code review bot(s) detectedCodeRabbit has detected other AI code review bot(s) in this pull request and will avoid duplicating their findings in the review comments. This may lead to a less comprehensive review. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Summary of ChangesHello @tabu-a, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed! This pull request integrates the HAMI project into the Kubernetes application, providing a robust solution for managing and scheduling heterogeneous AI computing resources. It introduces a new addon that streamlines the deployment and configuration of HAMI's core components, including its scheduler extender, device plugin, and admission webhook. This integration significantly enhances the platform's ability to handle diverse GPU and XPU hardware, offering flexible resource allocation and optimized workload placement for AI and high-performance computing environments. Highlights
🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console. Using Gemini Code AssistThe full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips. Invoking Gemini You can request assistance from Gemini at any point by creating a comment using either
Customization To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a Limitations & Feedback Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here. You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension. Footnotes
|
There was a problem hiding this comment.
Code Review
This pull request integrates the HAMi project for GPU virtualization by adding a new HelmRelease, updating configurations, and vendoring the upstream HAMi Helm chart. The changes are a good starting point for the integration. My review highlights a critical issue with a hardcoded node name in the vendored chart which will prevent it from working correctly. I've also identified a high-severity issue regarding the use of infinite retries in the HelmRelease, which could impact cluster stability. Additionally, I've provided some medium-severity suggestions to improve code style and consistency.
| { | ||
| "nodeconfig": [ | ||
| { | ||
| "name": "m5-cloudinfra-online02", | ||
| "operatingmode": "hami-core", | ||
| "devicememoryscaling": 1.8, | ||
| "devicesplitcount": 10, | ||
| "migstrategy":"none", | ||
| "filterdevices": { | ||
| "uuid": [], | ||
| "index": [] | ||
| } | ||
| } | ||
| ] | ||
| } |
There was a problem hiding this comment.
This ConfigMap contains a hardcoded node name m5-cloudinfra-online02. This will cause the device plugin configuration to be incorrect on any other node, preventing it from functioning correctly. This appears to be from an example configuration that was vendored. This hardcoded value should be removed. If no node-specific configuration is needed, an empty nodeconfig array would be a safer default.
{
"nodeconfig": []
}There was a problem hiding this comment.
looks irrelevant - this is a predefined value by hami maintaners
| install: | ||
| createNamespace: false | ||
| remediation: | ||
| retries: -1 |
There was a problem hiding this comment.
The remediation strategy for installation is set to infinite retries (retries: -1). This can mask underlying persistent issues and cause the Flux controller to enter a perpetual reconciliation loop, consuming significant resources. It is recommended to use a finite number of retries (e.g., 5) to handle transient errors while ensuring that persistent failures are reported.
retries: 5| upgrade: | ||
| force: true | ||
| remediation: | ||
| retries: -1 |
There was a problem hiding this comment.
The remediation strategy for upgrades is set to infinite retries (retries: -1). This can mask underlying persistent issues and cause the Flux controller to enter a perpetual reconciliation loop, consuming significant resources. It is recommended to use a finite number of retries (e.g., 5) to handle transient errors while ensuring that persistent failures are reported.
retries: 5|
|
||
|
|
There was a problem hiding this comment.
There are trailing empty lines with whitespace at the end of the file. These should be removed to improve code cleanliness and consistency.
| - name: {{ . }} | ||
| {{- end }} | ||
| {{- end }} | ||
| {{- end -}} No newline at end of file |
There was a problem hiding this comment.
This file is missing a newline character at the end. It's a common convention and good practice to end files with a newline to prevent issues with certain tools and to adhere to POSIX standards. Several other new files in this chart have the same issue.
{{- end -}}
Signed-off-by: Tamerlan Abu <tamerlanabu@gmail.com>
Signed-off-by: Tamerlan Abu <tamerlanabu@gmail.com>
No description provided.