The responses suggested verify and salvage. salvage(8) is not
available until 4.0D, but I managed to get it. Use of scu to verify
the disk was also suggested. Scu verify media reported no problems.
ADVFS verify looped on the same record and salvage memory faulted
(sctrace output makes it look to me like salvage is looping in the same
way that verify does). Since the behavior is the same for all three
disks (altough the looping with verify occurs at different places on
all of the disks) and I am only having actual problems with one of them
(and really only an advfs problems, it seems, vs. some hardware
problem), I don't think any of the disks are bad. My problem is not the
data from the one disk (I have that on tape), but to be certain that
when an advfs partition goes to lunch that I actually do have to
restore and that there isn't some other means to recover (I have had
advfs partitions go out twice before). Unless someone can explain the
weird behavior of verify and salvage, I will assume that that is what I
must do.
Thanks for any further clues and many thanks to the following
for their responses:
Dr. Tom Blinn <tpb_at_zk3.dec.com
Alan Rollow <alan_at_nabeth.cxo.dec.com
Larry Griffith <larry_at_garfield.wsc.mass.edu
John P.Speno <speno_at_isc.upenn.edu
Knut Hellebų <Knut.Hellebo_at_nho.hydro.com>
George Guethlein <GGuethlein_at_GiantOfMaryland.com>
John Becker <John.Becker_at_digital.com>
_____________________________________________________________________
/ Brian C. Hill bchill_at_bch.net *82 (530) 753-0358 \
| Systems Programmer University of California, Davis |
| Unix Specialist BCH Technical Services |
Received on Sat Jan 24 1998 - 04:04:19 NZDT