I've had a very strange corruption on a volume : all the files & folders are here, and all have 0 bytes, in lieu of the TBs of data it should contain.
Usual solutions (diskUtility repair, fsck, DiskWarrior) do not solve the problem, as far as the system is concerned the system is "clean".
Luck is: I have all the file inode numbers (ino) and file sizes, stored in a sqlLite database that I had created doing media-management.
Question: can I use this file size & ino info to do some manual restore? (for instance with dd from /dev/rdiskXsY)
Two side notes:
it's expected the filesystem to be not at all fragmented, as it's a copy-once / read-many use types, and was recently formatted
my suspect for this corruption is a third party had used an HFS+ driver from windows to create the data; but at one time, in the past, this data was fully readable, not corrupted*