giguere@csg.UWaterloo.CA (Eric Giguere) (07/04/90)
Here's the situation: we've set up a TCF cluster of PS/2's running AIX 1.2. So far I'm quite pleased with the performance. I would like to start regular backups, of course -- especially since the primary site has 200 megs of stuff on it. Obviously I don't want to backup to diskette! Our Sun 3/60 (running SunOS 4.0) however has an Exabyte 8mm tape drive (you know, the 8mm video tapes -- can hold up to 2 gigabytes each) which seems perfect. But when I try a backup using a sequence: rdump 0unf csg:/dev/rst9 /dev/hd2 where "csg" is the Sun's name and "/dev/rst9" is the tape drive off the Sun, the following happens: 1. The Exabyte writes stuff to the tape for a few seconds 2. The tape drive then seems to reset and rewind 3. Now the Exabyte starts writing the main data The rewind concerns me because I don't see why it's happening. What's more, typing rrestore tf csg:/dev/rst9 shows that the data written is corrupted -- all that shows up are lines of the form: Checksum error 0, inode 0 file <name unknown> This is very annoying, obviously! Then I got to wondering -- "rdump" is just a version of "dumpbsd"... however the Sun is most definitely not a BSD machine. Could this be my problem? I'm looking for solutions. I would prefer to keep the Exabyte drive on the Sun since I have to back it up as well and I don't feel like moving the thing from one machine to another all the time. Eric -- Eric Giguere giguere@csg.UWaterloo.CA