Removing big files is not possible (or very slow)

When a file should be purged we still create a hash. Once that is done we will actually try to back up the file to the orchestrator.

This needs to be optimized.

• Don't hash on read if purged=true
• Don't backup on delete/update