> Then I'd image the tape with Tapeimgr and see if I could figure out > what software was used. I've reverse-engineered a few, Retrospect and > one other, I think I uploaded it all to my Github.
Phil's overall process is good, but you shouldn't use `dd` to image tapes. It discards tape block length information, which is Bad <tm>. This Tapeimgr tool just uses dd under the covers. (In some scenarios, it _may_ be true that the tape blocks are all fixed length, but you may only get one pass through the tape, so why risk it?) https://github.com/brouhaha/tapeutils is probably a better choice. It does not do as much work with metadata, but you can collect and record such metadata in a sidecar file with no more effort than using Tapeimgr. For QIC carts and the like, the little rubber bands often go sticky, and may need to be carefully replaced. And definitely do testing of the drive with a scratch monkey before risking the irreplaceable target ones. De