This tutorial describes how to do automated server backups with the tool rdiff-backup. rdiff-backup lets you make backups over a network using SSH so that the data transfer is encrypted. The use of SSH makes rdiff-backup very secure because noone can read the data that is being transferred. rdiff-backup makes incremental backups, thus saving bandwidth.
The article states the use of SSH makes rdiff ‘very secure’. As always with this kind of statements, it should be taken with a grain of salt.
You could ask yourself if you think it’s a good idea to enable public-key authentication for root on all the machines you want to backup.
Also, at first glance it doesn’t look like the backed-up data is stored encrypted.
Nonetheless this seems like a nice article about good and useful tool, just some things to be aware of.
“Also, at first glance it doesn’t look like the backed-up data is stored encrypted.”
True. You want Duplicity for encrypted backup. This is a philosophical difference. In rdiff-backup’s author’s word: “rdiff-backup’s archives are meant to be as easy to view as possible.”
More on this here:
http://www.nongnu.org/rdiff-backup/duplicity.html
Ah, indeed it’s quite prominently linked on the rdiff page.
Great, linking to alternatives and explaining the when they might be a better idea is a Good Thing. Very mature/professional.
server backup-needs. it has already saved my ass several times:
http://folk.uio.no/johnen/bontmia/
i use it to do backups for multiple server. bontmia keeps historic versions (i.e. nightly for the last 14 days + weekly for the last 6 month + montly forever) but uses hard-links, so that only changed files use space on your harddisk.
Looks pretty swift, I’ll give it a try, thanks. I think the author should re-evaluate tape drives, they have been evolving, AIT for instance, more reliable hauling them off-site than throwing hard disks around and the speed I get is very, very good.
How is this different from rsync?
sounds exactly the same to me.
You rsync 2005-10-16 data to 2005-10-15 data backuped. Where is 2005-10-15 data now? How do you restore older copies?
You just copy backup before rsync? How is that an incremental backup then? You still waste disk space, although you conserve network bandwidth thanks to rsync.
cp -l
[quote]You rsync 2005-10-16 data to 2005-10-15 data backuped. Where is 2005-10-15 data now? How do you restore older copies? [/quote]
Actually if you were to look into rsync –link-dest=DIR you’d be presently surprised to find your answer.
rdiff-backup from my understanding is just an implementation of rsync. Mind you I haven’t read the article. I am simply stating.
rdiff-backup is ace. I’ve written a small backup tool around it, including an “interactive” restore module that tells me what backups I have, which increments are available, and asks where to restore them to. I run rdiff-backup nightly, so I can now restore the crucial parts of my filesystems to their state as of any of the previous X days.
I use subversion to back up small set of data, especially system config. I can go back to any previous date. If there is no change, it does not commit anything to the repository. Works great for me.
I’ve been using rsnapshot. It’s simple, written in perl, incremental, and uses hard links to save disk space. Oh, and it’s highly configurable!
I’ve been using rsnapshot as well. It works well and I really like it.
I have been using BackupPC at home and it is great.
http://backuppc.sourceforge.net