[comp.unix.ultrix] disk-to-disk dump fails on DS5800/DS5400

karrer@bernina.ethz.ch (Andreas Karrer) (02/08/91)

On our DECsystem 5840 and 5400, both equipped with RA90 disks and running
Ultrix 4.0, we used to do disk-to-disk backups of the root filesystem using 
the following script:

	#!/bin/sh
	BACKROOT=ra1a
	ROOT=ra0a
	/bin/umount /dev/$BACKROOT >/dev/null 2>&1
	/bin/newfs /dev/r$BACKROOT ra90
	if [ $? = 0 ]; then
        	if [ ! -d /backroot$$ ]; then
                	mkdir /backroot$$
        	fi
        	mount /dev/$BACKROOT /backroot$$
        	cd /backroot$$
        	/bin/dump 0f - /dev/$ROOT|/bin/restore rf - > /dev/console 2>&1
        	cd /
        	umount /backroot$$
        	rmdir /backroot$$
	fi

The partition sizes of /dev/ra0a and /dev/ra1a are exactly identical.

This has suddenly ceased to work. Now we get lots (15+) error messages
(from restore):
	expected next file 4194, got 4193
	expected next file 4205, got 4204
	expected next file 4219, got 4218
	...
	cannot find directory inode 11
	abort? [yn]

Am I missing something? 

+-----------
  Andi Karrer, Communication Systems, ETH Zuerich, Switzerland
  karrer@bernina.ethz.ch

schedler@src.dec.com (Richard Schedler) (02/20/91)

>	/bin/dump 0f - /dev/$ROOT|/bin/restore rf - > /dev/console 2>&1

Try using "restore xf -" instead of "restore rf -".  I don't know if
this is your problem, but it's always worked for me.

	--Richard
-- 

Richard Schedler                          Systems Research Center
schedler@src.dec.com                      Digital Equipment Corporation