Help with a backup solution

[BassTeQ2]

Hi all,

I have an idea for a backup solution, just like to see if it would actually work.

Ideally I'd like to use my 2nd HDD as a full drive backup incase the 1st HDD dies. Then the 2nd one can replace the 1st.

So I was thinking if I first do a full drive clone to my 2nd HDD, then use rsync to keep it updated.

my 1st drive looks like this

/dev/sda5    7.8G 917M 6.5G 13% /
/dev/sda8    645G 1.4G 611G 1% /home
/dev/sda6    996M 35M 910M 4% /tmp
/dev/sda3    7.8G 3.9G 3.6G 52% /usr
/dev/sda2    9.7G 431M 8.8G 5% /var
/dev/sda1    122M 18M 98M 16% /boot
tmpfs     1.9G  0 1.9G 0% /dev/shm

If I have cloned the 2nd hdd, would I then mount it as /backup ?
I'm kind of confused, could I have /backup/home, /backup/tmp, /backup/usr etc...

Each of these would link to its respective partition on the backup drive.

Does that make any sense :P ? Anyone have any ideas good/bad about this concept?

Thank you

[graeme_p]

You can use rsync and a perl script to keep rotating snapshots with minimal duplication.

You can use rsync for the initial copy as well.

I would also suggest using a USB external drive. That way you will not have both drives corrupted by a bad controller, or simultaneously trashed by a power surge or motherboard fault.

This is what I do, and I also do a network backup. I use the grsync GUI front end for rsync for both.

[BassTeQ2]

Thanks for the reply!

Would I need to keep all those partitions in sync? Or just perhaps /home, /usr, /var ?

Thanks

[graeme_p]

Depends on your requirements.

For desktop use I only keep /home in sync. For a server you might want to keep other directories in sync as well. Where is your (valuable) data? Do you have time to do a re-install to recover, or do you need to be able to swap drives and keep going?

I only do this for desktop use so I cannot confirm how well it works if your requirements are more extensive,

[BassTeQ2]

Its actually for a server, I'd like to be able to simply replace the drive and continue on.

[graeme_p]

It should work AFAIK, but you can probably find articles on this on the net, and relevant threads on the rsync list archice.

[Jamier101]

On my server I have two internal SATA drives, the first is partitioned and takes the operating system and primary copy of my work which is mount as work. I then formatted the second drive and mounted it as backup. Crontab then runs a backup script each night to ensure that I have a backup for everyday of the week therefore I have 7 days to realise a mistake.

In addition, I also have a LACIE USB hard-drive which acts as a removable copy which I can take off-site. The backup script e-mails me each morning to tell me if the backups were successful so that I can address any issues, the most common is that the LACIE has been powered down for one reason or another.

Although there is no way to remain 100% covered for everything I find the best policy is to change the drives in my server every 3 years or if any signs of bad sectors start appearing.

[graeme_p]

@Jamier101, do you use SMART disk monitoring? I just started when I realised that everything is pre-installed or in the repos - even nice GUIs to do it with on desktops.