With reference to case https://support.mongodb.com/case/00918578 we experienced situation where the migration was in a stuck state eternally. If the validation checks fail, then it is only fair and logical to fail the migration with a relevant error rather than the status lying about your shard migration is in progress.

Also, emit status updates on the migration, such as pre-validation checks passed, syncing data, and % complete on data migration or initial sync rather than just a green bar, and whatever else such as if it is building indexes, etc during migration. Basically, make the status as readable and logical as possible, and please introduce a timeout if the data migration isn't going on for real. I understand that you can't timeout since the time on migration can depend on the size of data but try to make the migration intelligent. It is very dumb atm.