Salutations to the Gurus of the 84,000 UNIX realms.
How would a UNIX bhiku or bhikuni of noble parentage resolve the
following problem in maximum harmony with the UNIXverse?
System disks from multiple AS/2100's of both EV4x and EV5x CPU types
are being backed up over a network to a Linux box disk drive using gtar
through rsh.
Under DU 3.2G the procedure for restoring the system disk was:
boot the system to be restored from an external scsi "rescue"
disk,
set up and partition the new disk with disklabel,
restore the files to the new system disk through the network
using gtar through rsh,
and recompile the kernel.
The AS 2100 being restored to may or may not be same one that the backup
was taken from. Actually we would like the restore procedure to be able
to restore a backup to ANY Alpha, whether it is a 2100 or not.
We have discovered (the hard way) that under 4.0D if the installed
EISA/ISA cards are different it may be necessary to run ECU prior to
recompiling the kernel. We suspect that because specific hardware
configurations are now being loaded from the sysconfigtab file under
4.0D that were previously loaded from the environment under 3.2G that
additional steps may be necessary to obtain a functioning machine.
Does anyone know how to force a regeneration of the sysconfigtab file?
Can anyone think of any other steps that need to be taken in the restore
process? Any comments on the efficiency of our overall backup strategy?
Namaste,
Gene
- application/ms-tnef attachment: stored
Received on Tue Sep 22 1998 - 22:48:36 NZST