Ask HN: Offsite backup, BUT without local copy of the data (with catalog)? | Hacker News Reader

Ask HN: Offsite backup, BUT without local copy of the data (with catalog)?

3 pointsdimitry123 comments

The problem: if I have 1TB of data, I need to have 1TB of "local" storage to make incremental backups with most tools. That consumes space and disables me to store my incremental backups encrypted/signed/etc

There is obviously the not-so-space-technology of backup catalogs. For windows systems there are: backup4all and Genie-Soft Backup Manager that use those.

For linux I could not find one. The closest alternative I could come with is 'git add ./*', 'git-gc' and then storing the packfile offsite, but it's a bit unstraight method, I suppose.

Are there any free/OSS alternatives for both linux and windows, which you can tell me about?

View on HN

I know this thread is old, but I ran across it in a search. I use a tool called dar (http://dar.linux.free.fr/) that is along the lines of tar, but with extra features for dealing with multi-volume backup sets and diffs easily.

It lets you extract a catalog from the backup set that you can use locally as a checkpoint for doing diff/incremental backups. The catalog is really small, mine is ~10Mb for hundreds of thousands of files.

  ## here is the initial backup: backing up /nas to
  ## an archive set /mnt/usb/nas on a usb device
  dar -R / -c /mnt/usb/nas -g nas

  ## extracting the catalog from the backup and 
  ## saving it under my home directory      
  dar -A /mnt/usb/nas -C /home/dan/backup/nas-catalog

  ## I use the local copy of the catalog in my home folder 
  ## when doing diff backups to another mounted usb device
  dar -R / -c /mnt/usb/nas-diff -g nas -A /home/dan/backup/nas-catalog

I don't quite understand the use case. Do you want to backup data and then delete the local copy? I think that's more like archiving than backup, and it's not clear to me what "incremental" even means in that case. Or do you intend to keep your 1 TB of live data, but you don't want the backup tool to use any additional disk space?

Speaking of catalogs, I wonder if Duplicity is what you want.

The scenario is: keep my 1 TB of data and store backup completely off-site, without having local copy of it when doing increments (so the backup tool can not compare the original data and backup to find the changes, but need to consult catalog instead).

Reply on news.ycombinator.com