kubeflow / kubeflow/trainer

TrainJob may be recreated after a failed JobSet is deleted

Open
#3,737 4 comments 0 reactions 0 assignees View on GitHub
area/controller kind/bug
Dominant language
Go
Stars
2.2k
Forks
1.1k
Avg merge
3d 22h
Merged PRs (30d)
39

Description

### What happened?

A terminally failed TrainJob may be restarted after its JobSet is deleted, either manually or by the JobSet TTL controller.

This can cause a failed training workload to run again unexpectedly and consume additional cluster resources.

### What did you expect to happen?

A terminally failed TrainJob should remain failed after its JobSet is deleted. Runtime objects should not be recreated unless an explicitly supported retry or recovery operation is requested.

### Environment

None

### Impacted by this bug?

Give it a 👍 We prioritize the issues with most 👍

Contributor guide

Open the contributing guide

Research direction

No files or tests are named. Start by tracing the TrainJob and JobSet reconciliation paths and reproduce deletion of a JobSet after terminal failure. Done means a terminally failed TrainJob stays failed and its runtime objects are not recreated without an explicitly supported retry or recovery operation.

Written by the indexing model from the issue text.

Assessment

Tech stack
go, kubernetes
Domain
distributed-systems, infrastructure
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
50/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.