On this page (and all other pages I checked at random), Wikimedia OCR returns a 400 HTTP status and the error "Image URL must begin with one of the following domain names and end with a valid file extension: upload.wikimedia.org and upload.wikimedia.beta.wmflabs.org".
The returned JSON is:
{ "engine":"tesseract", "langs":["en"], "psm":3, "crop":[], "image_hosts":["upload.wikimedia.org","upload.wikimedia.beta.wmflabs.org"], "image":"\/\/upload.wikimedia.org\/wikipedia\/commons\/thumb\/e\/e5\/Ling-Nam%3B_or%2C_Interior_views_of_southern_China%2C_including_explorations_in_the_hitherto_untraversed_island_of_Hainan_%28IA_cu31924023225307%29.pdf\/page160-1450px-Ling-Nam%3B_or%2C_Interior_views_of_southern_China%2C_including_explorations_in_the_hitherto_untraversed_island_of_Hainan_%28IA_cu31924023225307%29.pdf.jpg", "uselang":"en", "error":"Image URL must begin with one of the following domain names and end with a valid file extension: upload.wikimedia.org and upload.wikimedia.beta.wmflabs.org" }
Since the image URL actually matches the whitelist (modulo the leading //) it's possible the returned error is itself a bug triggered by an underlying problem.
In any case, from the user side this one is kinda "UBN"-y.
Edit: Oh, I should add, first reported at s:WS:S/H#Error_from_the_OCR_tool??.