AI-Hypercomputer / AI-Hypercomputer/xpk

[slice] e2e tests still passing after removing health label

未关闭
#614 14 条评论 0 个 reaction 已指派 1 人 已被 @mbobrovskyi 认领 在 GitHub 查看
主要语言
Python
星标
193
派生
94
平均合并
24 分钟
30 天内合并 PR
1

描述

We annotate each replicated job in JobSet here: https://github.com/AI-Hypercomputer/xpk/blob/1d2ac6b4d1e896cbd0ffde3781e0067a6ffd2812/slice/internal/webhooks/jobset_webhook.go#L66

However, after deletion/modification of the label used in e2e test, the tests still pass. Remove this line: https://github.com/AI-Hypercomputer/xpk/blob/1d2ac6b4d1e896cbd0ffde3781e0067a6ffd2812/slice/hack/kind-cluster.yaml#L30

And observe e2e tests still pass. Kueue should not return topology assignments on such nodes for these replicated jobs, which require this NodeLabel

**Additional Info**
Modifying the [loop](https://github.com/kubernetes-sigs/kueue/blob/b44cf15a52b5c97d46d172dcf2432228bd03403e/pkg/util/testingjobs/jobset/wrappers.go#L79) which creates each replicated job to add a node selector:
```
jt.Spec.Template.Spec.NodeSelector = map[string]string{"hello": "world"}
```
Results in Kueue's e2e tests failing. So, it seems that Kueue/TAS is acting upon these node selectors in some way. This makes me think that this is more likely a Slice Controller issue.

贡献指南

打开贡献指南

评估

这个 Issue 还没有评估数据。

把新 issue 发到你的邮箱

精选适合新手参与的 GitHub issue 摘要。