New Text Works Not Displaying Most Pages In File In UV

[KevinJonesMeta]

Descriptive summary

Three examples of ingested files that only display first page in viewer on Staging.

1. https://staging.oregondigital.org/concern/documents/wm117p06c (File downloaded from prod, OCR'd, ingested on staging 3/23/26 local time)

2. https://staging.oregondigital.org/concern/documents/pk02c987r (File downloaded from prod, OCR'd, ingested on staging 3/24/26 local time)

3. https://staging.oregondigital.org/concern/documents/hq37vn705 (File live on prod, downloaded from prod and ingested on staging 3/25/26 local time)

4. Example that DOES DISPLAY in viewer: https://staging.oregondigital.org/concern/documents/3484zh01z (File downloaded from prod, OCR'd, ingested on staging 3/23/26 local time)

Recently ingested text works on prod have not had this issue yet.

Expected behavior

All pages in file display in viewer after ingest jobs complete.

Related work

Accessibility Concerns

[KevinJonesMeta]

This appears to happening on production as well. Same file in production Review Submissions is only displaying first 8 pages in UV, same as on staging. The files are coming from Oregon Digital, into Acrobat for new OCR, and back to Oregon Digital. Conflict between new and old OCR on file?
Prod: https://oregondigital.org/concern/documents/ms35tb924
Staging: https://staging.oregondigital.org/concern/documents/cf95jb673

[KevinJonesMeta]

I loaded 1-3 above via bulkrax with exact same behavior, including matching pages displayed. There seems to be something with the files. Testing fresh file that I didn't edit/OCR and it loaded and displays fine. Unless it is related to file size, which I doubt.

[KevinJonesMeta]

My latest test used a different text recognition setting in Acrobat and has larger file size. https://staging.oregondigital.org/concern/documents/pc289j13p

Opened on my computer, the modified file is 482 pages as seen below (same as original downloaded file).

On staging in viewer the file has 7 pages. Previous tests had thumbnail placeholders that never filled in. This file didn't do that.

[KevinJonesMeta]

I reloaded General Catalog 1901-1902 downloaded from prod onto staging. Behavior is the same as the modified files.
https://staging.oregondigital.org/concern/documents/cj82k742g

[KevinJonesMeta]

General Catalog 1901-1902 after removing OCR and creating new OCR. https://staging.oregondigital.org/concern/documents/79407x30s

[KevinJonesMeta]

Based on my original attempt, I thought it was something to do with how I changed the OCR. So I've used two different OCR processes on the original file from prod, stripped out the old OCR and did new OCR, and loaded the unchanged file from prod on staging. In the case of General Catalog 1901-1902 none of them display past the first page on staging. Example links consolidated below for the same file. Behavior is the same in all cases.

Original from prod on staging: https://staging.oregondigital.org/concern/documents/cj82k742g
File after removing hidden information and redoing OCR: https://staging.oregondigital.org/concern/documents/79407x30s
Prod file after new OCR loaded via bulkrax: https://staging.oregondigital.org/concern/documents/p5547r481
Prod file after new OCR loaded via work form: https://staging.oregondigital.org/concern/documents/wm117p06c

1951-1952 file from prod using Enhance File -> Recognize Text in Acrobat: https://staging.oregondigital.org/concern/documents/pc289j13p [same behavior as above, just loads a few more pages first]

Counterexamples:
These two had new OCR added to file from production and display/are searchable on staging.
General Catalog 1950-1951 https://staging.oregondigital.org/concern/documents/rx913q03c
General Catalog 2004-2005 (from above) https://staging.oregondigital.org/concern/documents/3484zh01z

[morganem0]

@KevinJonesMeta, I checked the PDFs I deposited into OD-Staging March 13-19, and did not see this error in any of those.

These are PDFs ranging from 1 pg to over 100 pages, and in each PDF it looks like all pages are present, and I'm not seeing the 'broken-image' icon that I see in the example problem PDF you linked to here:
https://staging.oregondigital.org/concern/documents/hq37vn705

The PDFs I reviewed are in the ISSSSP Testing Collection here. I checked all 32 PDFs.
https://staging.oregondigital.org/catalog?f%5Bfile_format_sim%5D%5B%5D=pdf+%28Portable+Document+Format%29&f%5Bnon_user_collections_ssim%5D%5B%5D=test-ISSSSP&f%5Bresource_type_label_sim%5D%5B%5D=Text&locale=en&page=2

Let me know if there's other stuff to check, if I can help out with troubleshooting this.