SDAPI works with errors

[KatharineV]

Entering the catalog view of the SDAPI tenant, the first page of works contains several that will not open. The works generate error messages. In the catalog search, the works with PDFs attached are the ones with error messages. Citation-only works (without files) are opening fine. From the homepage of the SDAPI tenant, the featured works with attached files are opening without errors.
https://sdapi.b2.adventistdigitallibrary.org/catalog?utf8=%E2%9C%93&locale=en&search_field=all_fields&q=

The only thing that I have done with this tenant since the Valkryie cutover was that yesterday (12/4) I ran two importers to clean up sets of works we had tried to import before Valkyrie, but Fedora glitched and dropped hundreds of works. So I ran the importers yesterday and brought in the missing works. The importers did NOT run cleanly, or else the enqueued and failed counts on the dashboard failed to update. I don't know which. I will create a new ticket for those issues, either way. I can't confirm what's going on with the importers because they will not load. They spin for several minutes and never load and/or then I get a Cloudflare error.

These are the importers:
https://sdapi.b2.adventistdigitallibrary.org/importers/557?locale=en
https://sdapi.b2.adventistdigitallibrary.org/importers/171?locale=en

The importers that ran yesterday do NOT contain the works that are listed on page one of the catalog, because the importers yesterday did NOT contain works with attached files. So I am confused why there is a connection between what I did and why certain works with errors have bumped to the top of the catalog search (as though they are most relevant, which I believe generally means "most recently updated and/or uploaded). These errored works were NOT at the top of the catalog search prior to me running the importers yesterday. It seems connected. I can't say how.

[KatharineV]

2024-12-05.SDAPI_works_with_errors.mp4

[ShanaLMoore]

After pairing with Rob we've determined that this is an infrastructure related issue.

Rob will restart the database.