shorvath <email@example.com> wrote
on 06/26/2012 03:10:59 PM:
> Yes, I could just take the backup directly from the individual
> servers, and by the looks of things it may be my only option.
> I was rather hoping for perhaps another fuse module alternative or
> something that would still allow me to keep everything centralised.
> I really don't want to have backup scripts running on every server
> if I can help it.
Why would the backup have to be handled by the host?
You can use whatever technique BackupPC is using to pull the data,
all from a central location just like BackupPC.
If you were using rsync over ssh or rsyncd, it would
be quite trivial to create a system that looked just like BackupPC to the
hosts but instead of storing all files in a single pool and handling versioning,
dedupe, etc., it simply dumped the files into the filesystem directly.
If you're using some other method (tar or smb, for example), it would
still work (and trivially), but would be very hard on bandwidth. If
the data is being moved over a WAN link, performance will be an issue.
Stick with rsyncd or rsync over ssh.