Re: Strange error saving IFS objects (my "enhanced SAVCHGOBJ")

James H. H. Lampert wrote:

Preparing to do just that [saving to a SAVF first, then SAVSAVFDTAing

> the SAVF to the tape]. Certainly worth a try. And there are some old

QIC120 tapes around, on which it shouldn't be difficult to force a volume change.

My thinking is that it *should* work, since the OS bug is supposedly something to do with locks on the *objects*.

You wouldn't believe how long the experiments took. Using QIC150 tapes in order to hit a volume change as quickly as possible, I first found that I had an IFS subtree that was too honking big to fit on the 2 tapes I'd allocated to the project. Then, after using a smaller one that took 11 SAVSAVFDTAs to hit a volume change, I snapped a belt (and evidently jammed the drive wheel) on the second QIC tape. Finally, the third time around (using a QIC150 shorty as the first tape!) worked, and only took an hour.

I learned early-on that it solved the lock problem, but I needed a complete save that crossed a volume change in order to confirm that it would restore properly.

--
JHHL
(I started looking for a freshly painted wall to watch, while I waited for those damned saves to complete!)