[Virtual-pkg-base-maintainers] Bug#622633: base: single fixed disk is arbitrarily identified as /dev/sda or /dev/sdb, causing dumps to fail

[Dean Allen Provins]

Package: base
Severity: important


I have used "dump" and "restore" to perform system backups for many years.
Since upgrading to Debian 6.x, I have not been able to obtain consistent and 
reliable dumps for the following reason:

   Sometimes. my single fixed disk is labeled as /dev/sda, but

   At other times, it is labeled as /dev/sdb.

My "dump" script identifies a partition to dump by its name (i.e. /home)
or partition (i.e. /dev/sda3) depending on the dump level.

If /var/lib/dumpdates indicates that the last lower value dump was performed
on partition /dev/sda3, and the system has assigned that partition to
/dev/sda3, then all is well;  however if the system has
arbitrarily labeled that partition as /dev/sdb3, then "dump" thinks that no 
lower level dump was ever performed on that partition, and it attempts to 
perform a level 0 (i.e. full dump) dump and the tape in my tape drive is
insufficiently long to handle that amount of data.

This means that I must NOT rely on my automatic (crontab-based) dump
scripts, but interrogate the system manually, and if necessary, alter 
/var/lib/dumpdates so that the script will run properly.

This is a REAL PAIN.

Is it possible that /etc/fstab, which now identifies the partitions on my
single fixed disc via UUID labels, is an unwilling participant in this
confusion?

Should I alter /dev/fstab to indicate the partitions as it was done before
(i.e. /dev/sda1 is /, /dev/sda3 is /home etc.)?

I look forward to your analysis and recommendation.

Dean

-- System Information:
Debian Release: 6.0.1
  APT prefers stable-updates
  APT policy: (500, 'stable-updates'), (500, 'stable')
Architecture: i386 (i686)

Kernel: Linux 2.6.32-5-686 (SMP w/2 CPU cores)
Locale: LANG=en_CA.utf8, LC_CTYPE=en_CA.utf8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash