I am stuck where I have a backup server with some storage and > 30 servers. I can't due to space keep live rsync of the code, but wish to have backups of these box's. *Note this will be backed up to disk, not tape or other media*

I played with some scripts, I know tar has an --update, but still having a problem with space, then the extraction of multiple file locations for configs, websites, etc.

I have used bakula in the past, but not sure it will suit my needs. I can get additional storage if need be as well as drives for a 2nd box, but looking here for suggestions.

I need to backup large > 100GB of some things like mail, websites, database, etc. some of which will change, some won't. So it would be nice to have a master archive, then a --differential I think would be the ideal solution with minimal downtime (which is why an image type solution cant be used).

I await great ideas and suggestions!
Thanks

If you have that many servers, then it makes sense to have a backup server with enough storage (either as a disk array or a tape drive or library) to back them all up. I use Amanda with an AIT5 tape library. More common (especially now that Sony has EOL'd AIT) is LTO, though many people are going with disk arrays since very large drives are becoming cheap.

I wrote up the reasons I like Amanda, and they are still relevant, even though a couple of years have passed. You can get the latest from the Zmanda site and the user documentation from the wiki. You can use it to backup to tape, disk or the cloud.

1 members found this post helpful.

Thanks, I was reading numerous sites, and yes I do agree having enough strorage.

Total compressed is around 1T and the storage device has 1 1/2T available, so it is enough for a full + incremental. I am looking into a few storage array's, etc. but the 1st and formost is getting the right app in place.

I will take a look at your doc now as well to help in the decision making.

Thanks.

Look into rdiff-backup http://www.nongnu.org/rdiff-backup/ as well. A command line utility that basically does a rsync from a local or remote disk. Your left with a directory that is essentially a snapshot of the area at the time the backups ran. I share this directory out to the users, read-only, so they can complete their own recoveries (permissions are the same). In addition, rdiff-backup completes incremental backups as well. It only saves the changed data so the space needed would be smaller when compared to a "normal" incremental. I'll let the website explain the rest.