[with apologies to Marcel Proust]

When finalizing the copy of all contents to azure, 225 objects failed to copy for two reasons:

• file not found
• file empty

This means that, for all intents and purposes, these contents are lost.

A CSV export of the metadata for these contents is available here.

To try to restore these, we need to first find where they came from, to see whether they're still available at the source. If they are, we can ingest them manually. If they're not, for now I believe we're out of luck, unless someone has better ideas.

Update (2019-06-20): 74 contents are left missing