eBackup - Problems, questions, solutions some thoughts

[dzingis]

Well backup/restore problem from me again 

My opinion is that backup/restore module must the most stable one, and new one in 1.4 is a good one.

I was doing some testing on real machine using eBox 1.4 with backup and restore, and here are my conclusions.

HDDs on server are configured using the automatic mode, not expert, so LVM modules are created.

I was trying to rsync backup to Windows XP machine with no success. Someone did that? (i've installed OpenSSH, copied rsync to OpenSSH directory, but after that rsync starts to complain about some /usr folder and quits).

My winning combination was scp, but problem was that scp backups data in folder where is OpenSSh installed (it tries to mount /home/username folder). That can be solved tweaking the client side passwd file in /etc folder (using cygdrive keyword).

Now I have my working backups Only problem I saw was that if you change the backups are done (scp, rsync, ftp) you loose connection to them in Restore files tab.

Now comes the magic.

My scenario was:

1. Server is dead
2. I have spare one but with disk off different size. (smaller)
3. I'm doing full restore with one full and one incremental backup.

I've followed steps from manual, and first thing to mention is:
- If you use use full image of grml you don't need to install duplicity because duplicity is included, why I'm saying this? If you do apt-get install duplicity it will install, and make no complains but you wouldn't be able to make restore because some silly error with temp folder

After that I've followed some steps from:

Code: [Select]

http://www.linuxforums.org/articles/lvm-baremetal-restore-step-by-step_397.html
with no luck because my disk was smaller, then I took exact the same one with same result.

What I've done:

1. copied MBR from "old" disk to "new" disk
2. did mkfs on /dev/sda1 and /dev/sda5
3. restored /boot/ directory to sda1
4. restored everything else in sda5
5. after reboot, I got an error 15 from grub that was solved recreating grub using Ubuntu LiveCD (does grml have solution for that ?)
6. after that I restarted server, and it started to complain about /dev/mapper directory that is missing (we deleted it in the last step of manual). Would restoring that directory solve the problem?
7. What about UUIDs in /etc/fstab file after restore?

Those UUIDs could be the problem with the plain install (everything on one disk). Am I right?

I will continue testing tomorrow and let you know. (if someone is interested )

I hope someone understood what I was trying to say and also hope someone tried to do same as me and succeeded  

regards