con / con/mechababs

ds006623: annex content not retrievable — never mirrored to s3-PUBLIC (only copy on internal OpenNeuro remote)

Open
#43 1 comment 0 reactions 0 assignees View on GitHub
dataset upstream
Dominant language
Python
Stars
1
Forks
4
Avg merge
15h 39m
Merged PRs (30d)
24

Description

**TL;DR:** ds006623's anat job fails because `datalad get` of its imaging files returns `not available`. `git annex whereis` shows the only copy is on the internal `OpenNeuro` repo (`185fe801`); the public `s3-PUBLIC` sibling (what our jobs fetch from) holds **zero** copies, dataset-wide. Reproduced off-cluster on a fresh clone → **not transient, not a dartfs/ACL issue**. The content appears never to have been mirrored to `s3-PUBLIC`.

## Minimal reproducer

```
datalad clone https://github.com/OpenNeuroDatasets/ds006623
cd ds006623
git annex whereis sub-02/anat/sub-02_T1w.nii.gz
# whereis sub-02/anat/sub-02_T1w.nii.gz (1 copy)
# 185fe801-93f0-42d1-b387-a3568a0b374a -- OpenNeuro
datalad get sub-02/anat/sub-02_T1w.nii.gz
# get(error): ... [not available; (Note that these git remotes have annex-ignore set: origin)]
```

`s3-PUBLIC` exists as a sibling but holds no content; `s3-BACKUP` doesn't have it either (so enabling it wouldn't help). Sampling other subjects' T1w shows the same — only on `OpenNeuro`, dataset-wide.

## Action
Raise with OpenNeuro: the dataset's content isn't retrievable via the public path (`s3-PUBLIC` unpopulated). Likely via Yarik. Drop `upstream-NOT-FILED` once filed.

## Open (minor)
Whether the `OpenNeuro` remote can be enabled as a stopgap — but the real fix is upstream.

## Done when
ds006623's content is retrievable via `s3-PUBLIC` (or we accept it as a known-unavailable dataset).

Contributor guide

No contributing guide indexed for this repository

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.