The following thumbnail image URL is being served up though a fully integrated Islandora/Fedora Commons subsystem:
http://goldmine.uncc.edu/islandora/object/motorsports%3A1/datastream/TN/...

Watchdog reports:
Nested HTTP request to generate islandora/object/motorsports%3A1/datastream/TN/view: failure (URL: http://goldmine.uncc.edu/islandora/object/motorsports%3A1/datastream/TN/..., time: 1467304036).

CDN module appears to be trying very hard to do its thing, but failing. How do I approach going after the problem?

The only clue I've been able to find is Mr. Leers' commentary about using file_create_url():
http://wimleers.com/comment/1501#comment-1501

Any assistance and thoughts would be very much appreciated, thank you ahead of time.

Comments

bradspry created an issue. See original summary.

wim leers’s picture

Status: Active » Postponed (maintainer needs more info)

Is http://goldmine.uncc.edu/islandora/object/motorsports%3A1/datastream/TN/view the actual image URL? That doesn't look like an image URL at all.

bradspry’s picture

Yes Sir, that's an Islandora-specific image URL to render thumbnail images.

Here is a collection page full of rendered thumbnail images:
http://goldmine.uncc.edu/islandora/object/motorsports%3A1949?page=1

Thank You For Your Time!

wim leers’s picture

What is responsible for that URL? I don't see an image extension. It looks like that URL streams an image, it doesn't generate an image? Is this perhaps some private file system-esque thing (private://)?

bradspry’s picture

Yes, streams an image is accurate and the best way to describe it.

Islandora's Tuque library is ultimately responsible, it makes curl calls to Fedora Common's REST-based web interface:
https://github.com/Islandora/tuque/blob/fd041d3a4a50a708ab676116329a7613...

https://wiki.duraspace.org/display/FEDORA38/REST+API#RESTAPI-getDatastream

wim leers’s picture

So it's streaming data. The CDN module assumes that any file that doesn't yet exist on disk will exist on disk after it's been requested one time. That's not the case here.

These then are not file URLs: they don't link to files that exist on disk.

There are a few options here:

  • generate these file URLs by using a custom stream wrapper
  • generate the URLs using the url() function: then the CDN module won't touch them at all
  • tell the CDN module to ignore all islandora/objects/* URLs, by blacklisting them in the CDN module's settings.
wim leers’s picture

Status: Postponed (maintainer needs more info) » Fixed

Status: Fixed » Closed (fixed)

Automatically closed - issue fixed for 2 weeks with no activity.