Like this one (has DjVu and PDF, but no JP2): https://archive.org/download/youthfulchristi00unkngoog
This is recorded in the logs as LOG.CRITICAL: Zip file not found.
We should check for the existence of the zip file before allowing people to save the job to the queue. Or, better: we could look for other images such as _tif.zip and extract accordingly.
This probably only affects older IA items, because I think the modern system is to create a standard JP2 zip for everything.