CI error: `All chunks should've been deleted` keeps coming back

## 🐛 Bug

https://github.com/Lightning-AI/litdata/actions/runs/12377277308/job/34546400027

This CI issue keeps coming back, even though it was fixed before.

### To Reproduce

Steps to reproduce the behavior...

Code sample

```python
# Ideally attach a minimal code sample to reproduce the decried issue.
# Minimal means having the shortest code but still preserving the bug.
```

### Expected behavior

### Additional context

Environment detail

- PyTorch Version (e.g., 1.0):
- OS (e.g., Linux):
- How you installed PyTorch (`conda`, `pip`, source):
- Build command you used (if compiling from source):
- Python version:
- CUDA/cuDNN version:
- GPU models and configuration:
- Any other relevant information:

Start with the linked GitHub Actions job and inspect the `All chunks should've been deleted` failure. The issue names no repository file, test, reproduction steps, or expected behavior, so trace the failing CI path and any earlier fix before choosing a change. Done means reproducing the failure and getting the affected CI job to pass reliably.

Written by the indexing model from the issue text.