Skip to content

Import failure: on005279 (ds005279) #97

Open

Description

@nemar-bot

Dataset: on005279
Source: OpenNeuro ds005279
Stage: prepare
Error: terminal: prepare=failure copy=failure finalize=failure
Workflow run: https://github.com/nemarDatasets/.github/actions/runs/32210575391
Hint (best-effort, from stage=prepare, not authoritative): possible git-divergence

See nemarOrg/nemar-cli#967 and docs/import-failure-procedure.md for the triage procedure.

Activity

  1. added
    needs-triageNo cause could be determined from the reported error; classify by hand from the run
    on Sep 11, 2026
  2. nemar-bot commented on Sep 11, 2026

    @nemar-bot
    Author

    Relabelled (2026-09-11T03:00:33.883Z): cause changed: (none) -> needs-triage.

    No cause could be determined from the reported error. Open the workflow run to classify it by hand.

  3. neuromechanist commented on Sep 16, 2026

    @neuromechanist
    Member

    Re-measured: above the 90% threshold, stays listed (#1396, ADR 0064)

    Re-measured against the bucket on 2026-09-16, on ADR 0064's basis (distinct
    annexed keys, never tree entries and never total_files):

    on005279: 30 of 1248 data keys have no object. 97.6% of its data is available.

    Above the 90% a listed dataset must reach, so ADR 0005 governs unchanged: the
    dataset is listed, serves what it has, omits what it does not, and advertises
    nothing it cannot deliver. No withdrawal, and the remaining gap is reported
    rather than hidden.

    Thirty keys of 1,248, and this dataset was recovered and reinstated on
    2026-09-15: it was one of six of the eleven withdrawn datasets that turned out to
    be fetchable by OpenNeuro own advertised route the whole time they sat private
    with tombstoned DOIs. The remaining 30 are anatomical images OpenNeuro deleted;
    their false presence claims were retracted as part of the 230 keys across five
    datasets, and verified from a fresh clone of origin. At 97.6% it is above the
    threshold, so it stays listed with the gap declared. This is its final state
    unless upstream restores those images.

    Verification basis: presence at declared size, so a zero-byte object left by
    a failed copy counts as missing rather than present (#967). Size-level, not
    checksum-level.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    import-failureOpenNeuro import/recovery failed for a datasetneeds-triageNo cause could be determined from the reported error; classify by hand from the run

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions