Network Backup

[nkjha]

Hi all,
Now I am using Fedora 7 and Fedora 11 on my all network machine.
I want to take the backup of all my Linux machine on a central linux machine.
Which backup option is best for me .


Thanks,
Nkjha

[Mudgen]

My money's on rsync as the core mechanism. What are your retention requirements?

[Rubberman]

I use rsync for this also. See the man page for details on its use.

[aber0cks]

I find it shocking that in nearly all forums/websites everyone suggests rsync. Surely there are alternatives?

I know you said fedora but duplicity is quite good (works on debian not sure about other OSs).

help.ubuntu.com/community/DuplicityBackupHowto#Restore.sh

I have used rsync but its more of a mirroring process rather than a historical backup process. Unless I'm mistaken?

[Mudgen]

I find it shocking that in nearly all forums/websites everyone suggests rsync. Surely there are alternatives?

I know you said fedora but duplicity is quite good (works on debian not sure about other OSs).


I have used rsync but its more of a mirroring process rather than a historical backup process. Unless I'm mistaken?

I don't see any historical backup requirements stated, and I inquired about retention requirements. Give simplistic requirements, get simplistic solutions. Refine them, and get refined solutions from the community.

[aber0cks]

Apologies, I'm not saying there's anything wrong with suggesting the mirroring route, it's just something to take into consideration

[Rubberman]

You can easily have rsync skip un-modified files/directories. It will only copy over the deltas between versions and has the option to compress the data during the transfer, making it a very good performing solution for networks. Anyway, if you read the documentation (man page) you will see that it has both archival and mirroring capabilities.

[aber0cks]

Actually after my little rant and looking at man pages, duplicity actually implements rsync LOL

[Rubberman]

Actually after my little rant and looking at man pages, duplicity actually implements rsync LOL

You can consider rsync a backup engine since it is a command-line only tool. Adding a decent UI, more user-friendly selection options, and the ability to make point-in-time backups for archival purposes, is an appropriate way to provide an enhanced experience and backup system tool. A major problem with many commercial backup tools is that they store the data in a proprietary format, so moving the archives to an incompatible system, or system the tool vendor doesn't support, can result in a lot of unusable backup data. Not a good situation as far as I'm concerned. This has bitten me and my clients many times in the past, which is why I much prefer an open source solution for such critical computing infrastructure components. At least a GUI wrapper and such doesn't keep me from accessing the data when/where I want/need.

[Mudgen]

As usual, Rub, you hit the nail squarely on the head. I've found myself too many times with backup tapes, status messages that say the backups completed, and a busted gui that refused to restore from them. These days I want something I know how to get under the hood of. rsync stores the files in their native format, and you can get to one or all readily.

As to historical retention, in my opinion that's a something to be implemented on the target backup server, with offsite rotation of media or offsite mirroring. I like to keep the client side simple, just rsync to a target.