Skip to content

hurlimage.soest.hawaii.edu images unreachable — infinite 301 redirect loop (1,730/6,463 train images) #1

@z3301

Description

@z3301

All images hosted on hurlimage.soest.hawaii.edu are currently unreachable due to a misconfigured nginx redirect. The server returns a 301 that prepends /imagearchive/ to every request, creating an infinite redirect loop:

$ curl -sI "https://hurlimage.soest.hawaii.edu/SupplementalPhotos/Localization_Scenes/D2-EX2301-06/Deep_Discoverer_23010006--6dba2f1b-4aa0-41c9-bdff-fa61343f30a9.jpg"

HTTP/1.1 301 Moved Permanently
Server: nginx/1.29.5
Location: https://hurlimage.soest.hawaii.edu/imagearchive/SupplementalPhotos/...

$ curl -sI "https://hurlimage.soest.hawaii.edu/imagearchive/SupplementalPhotos/..."

HTTP/1.1 301 Moved Permanently
Location: https://hurlimage.soest.hawaii.edu/imagearchive/imagearchive/SupplementalPhotos/...

Each subsequent request adds another /imagearchive/ prefix, looping indefinitely.

Impact

  • 1,730 of 6,463 training images (26.8%) are affected — all URLs with hurlimage.soest.hawaii.edu host
  • 3,369 of 22,225 annotations (15.2%) are inaccessible
  • Several categories are severely impacted:
Category Missing Total % Lost
benthic worm 60 60 100%
jelly 114 114 100%
amphipod 54 58 93%
sea pen 41 50 82%
bivalve 44 56 79%
sea cucumber 289 378 76%
sea snail 30 52 58%
anemone 571 1124 51%
sea star 319 624 51%
isopod 14 28 50%
  • The 4,733 images hosted on d2l7vcm2vanphr.cloudfront.net download without issue.

Environment

  • Tested with curl, httpx (via download.py from this repo), and the FathomNet API
  • The FathomNet API (fathomnet-py) also returns the same broken URLs
  • The issue is server-side (nginx misconfiguration on hurlimage.soest.hawaii.edu)

Request

Could the server team fix the nginx redirect configuration, or could mirror URLs (e.g., via CloudFront) be provided for these images? The download.py script supports resuming, so participants can re-run it once the URLs are fixed.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions