search

scientist-softserv / adventist_knapsack

Apache License 2.0
1 stars 0 forks source link

Some works won't open #521

Open KatharineV opened 4 months ago

KatharineV commented 4 months ago

Some works won't open on ADL production. They are returned in search results, but clicking the work results in the page spinning to load for a long time, finally ending in a 504 Gateway Time-out.

EXAMPLE A https://adl.b2.adventistdigitallibrary.org/concern/published_works/22261317_the_bible_made_plain_a_series_of_short_bible_studies_for_the_hom

Image

EXAMPLE B https://adl.b2.adventistdigitallibrary.org/concern/published_works/22253157_spectrum_april_1976

Image

KatharineV commented 4 months ago

New upload from Tuesday was accessible right after we created the work. Now it spins, won't load, and ends in Gateway Timeout. Since this isn't just an isolated issue affecting older works that were uploaded when the site was in significant development, I thought I'd add this recent example.

https://adl.b2.adventistdigitallibrary.org/concern/theses/b1175307_a_study_of_the_contribution_made_to_the_seventh_day_adventist_move?locale=en&parent_query=b1175307

KatharineV commented 3 months ago

Adding some context I'm seeing with recent uploads (as of 2024-03-28). I ran a small importer yesterday and uploaded five works to ADL prod. The big uploads are paused, so I understood that the workers should be ready to go and not swamped. The import seemed to go well. It succeeded and the works appeared with their metadata and files pulled from S3 via remote_url. Files were displaying in PDF.js prior to splitting. The works were not yet loaded into their collection, which was designated in the "parents" column of the CSV. I could view all of the work pages, although they loaded slowly.

https://adl.b2.adventistdigitallibrary.org/importers/300?locale=en

Later last night and again this morning I tried to view the works. The pages will no longer load. They spin and fail with a 504 gateway timeout error. The works now appear in their parent collection, and the notifications page says that the batch create has passed for all five works. I do see child works in the dashboard, although I can't see them in the UV because the parent works won't open.

So what changed between viewing and losing ability to view the works?

  1. PDF splitting
  2. Collection relationships

Image

I hope this additional info helps.

KatharineV commented 2 months ago

Another update: I'm trying to open a work that was uploaded in July 2023, long before I noticed this issue happening with uploads. The work is timing out and won't open. I believe the work may have been uploaded through the UI, based on the identifier. But I am unclear. Still, it is inaccessible and would be unrelated to any of the recent ingests or uploads that resulted in inaccessible works.

https://adl.b2.adventistdigitallibrary.org/concern/published_works/b4273079_a_reply_to_the_shepherd_s_rod?

laritakr commented 2 months ago

The work is in fedora and has the file sets attached. The direct file set link works. https://adl.b2.adventistdigitallibrary.org/concern/file_sets/ca6ce3bf-2bf4-4656-b0ef-5942693b000e?locale=en

Reindexing the work took a very long time and did not resolve the issue. I am not finding anything in Sentry to explain what might be happening.