0

I have a drive with data on it. I dont want to use RAID, instead I want to do hourly backups to a second drive. I basically want to mirror the data drive and resync every hour.

It is inefficient to re-move the data each time so really I only want to move across what changes. I.e if I add a new file to the data drive only that file will be moved across.

What tools are there for doing this on the command line? I used to use luckybackup on Ubuntu but now im on commandline debian.

James Willson
  • 125
  • 2
  • 1
    Have you tried rsync? It can sync to a directory, and does some diffing to prevent a full re-copy. – ssube Oct 28 '12 at 14:57
  • Do you want multiple snapshots or just one? If it's just one copy then rsync should do the job. – DerfK Oct 28 '12 at 14:59
  • I just want one drive that is an exact copy of the other. I've read about rsync, but it looks a little complicated. I will do some more reading. – James Willson Oct 28 '12 at 15:10
  • Like this? rsync -az /media/tv/ /media/tvbackup/ – James Willson Oct 28 '12 at 15:17
  • +1 for rsync or any other **differential** backup program. And a second sentence to remind you that a copy on the same computer if not the same as a backup. :) – Hennes Oct 28 '12 at 15:38
  • What's wrong with a RAID1, where the backup is there almost instantly? – ott-- Oct 28 '12 at 18:58
  • Because I worry about my files being corrupted and then instantly copied as a corrupt file. On top of this my drives are full and I think raid needs an empty drive? – James Willson Oct 28 '12 at 19:08

2 Answers2

1

How about rsnapshot? It uses the mentioned rsync for the actual copying and allows definitions of different backup schedules etc. Plus it allows to roll back to a different snapshot if needed. It is also pretty space saving as it will create hardlinks and not copy files twice if an older backup of the same file exists.

Yes, you can achieve all of the above with rsync and some self-written scripts, but I do not see a need to do the work twice :)

BTW, hourly copies sound a bit often. Bear in mind that any solution you chose will have to compare ALL your files once an hour for changes, slowing down the rest of your work. I usually run backups at 3am in the night, when (hopefully) nobody uses the machine.

zhenech
  • 1,492
  • 9
  • 13
0

I use LV snapshots for something similar. It allows you to create instantaneous backups and only uses disk space to track between that point in time and the changes you are making. This means you can make lots of snapshots with very little disk requirements.

See: http://www.softpanorama.org/Internals/Unix_filesystems/snapshots.shtml

Alastair McCormack
  • 2,184
  • 13
  • 22