vllm-project / vllm-project/aibrix

Bug: StormSet scale-in does not prioritize deleting non-ready RoleSets

Open
#2,082 1 comment 0 reactions 1 assignee Claimed by @googs1025 View on GitHub
Dominant language
Go
Stars
5.1k
Forks
694
Avg merge
1d 19h
Merged PRs (30d)
98

Description

### 🐛 Describe the bug

**Description**

When performing a scale-in operation on StormSet, the controller does not prioritize deleting non-ready RoleSets. This may lead to suboptimal behavior, where ready RoleSets are deleted before non-ready ones.

**Possible Root Cause**
https://github.com/vllm-project/aibrix/blob/main/pkg/controller/stormservice/utils.go#L227

When i.revision == j.revision == updatedRevision, should check the readiness of the RoleSets[i]

### Steps to Reproduce

1. Create a StormService with 5 replicas.
2. Scale the StormService out to 6 replicas.
3. Scale the StormService back in to 5 replicas.

### Expected behavior

During scale-in, non-ready RoleSets should be deleted first to ensure system stability and better resource utilization

### Environment

AIBrix version: 0.5

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.