Redo Backup Notes

gistfile1.txt

The following are command line commands for backing up /dev/sda1 in the same manner 
that "redo backup and restore" uses.  Note that the MD5 of the partition files will
not be the same because gzip stores info about the current time 
(http://superuser.com/questions/617934/):

#mount network drive
mount.cifs '//host.example.com/fab/temp' /mnt/backup -o username=myuser,password=mypass

#save backup list (sda1 in our case)
echo sda1 > /mnt/backup/20141009.backup

#save MBR
dd if=/dev/sda of=/mnt/backup/20141009.mbr bs=32768 count=1

#save sfdisk info
sfdisk -dx /dev/sda > /mnt/backup/20141009.sfdisk

#save size of partition
BYTES=$(fdisk -l /dev/sda | grep '/dev/sda:' | perl -pe 's/.*, (\d+)\D.*/$1/')
echo $BYTES > /mnt/backup/20141009.size

#save partition
partclone.ntfs -c -F -L /partclone.log -s /dev/sda1 | gzip -c --fast | split -d -a 3 -b 2048m - //mnt/backup/20141009_part1. 


Here are the commands for doing a restore:

#restore MBR
dd of=/dev/sda if=/mnt/backup/20141009.mbr bs=32768 count=1; sync;

#restore partition table
sfdisk -fx /dev/sda < /mnt/backup/20141009.sfdisk; sync

#check partition table (redo verifies that output is "1")
file /dev/sda1 | grep 'block special' | wc -l

#get number of blocks for sda1
cat /sys/block/sda/sda1/size

#get number of bytes for sda1 (redo probably ensures this matches .size file)
echo $[ $(cat /sys/block/sda/sda1/size) * 512 ]

#zero out beginning of sda1 (why?)
umount /mnt/sda1 2>&1; dd if=/dev/zero of=/dev/sda1 bs=1K count=1000; sync

umount /dev/sda1 2>&1

#do the restore
cat //mnt/backup/20141009_part1.* | gzip -d -c | partclone.restore -F -L /partclone.log -O /dev/sda1