Colin, Yes, that makes perfect sense! Thanks for the prompt reply. I was unaware of the metadata cataloging Tarsnap does, so that should make subsequent backups very quick as you suggest.
Regards, Seth On 8/24/17, 3:35 AM, "Colin Percival" <cperc...@tarsnap.com> wrote: On 08/23/17 17:40, Seth P. Garrepy wrote: > The backup is a flat-file dataset containing assorted documents, videos, > and audio files. I am trying to assess my initial costs, since I started > with you guys very recently, before I send my initial backup over the wire; > I am running it in a dry-run mode at the moment. The raw size of the > dataset is 2.3TB running on a RAIDZ2 volume consisting of 4x4TB WD Red NAS > drives (SATA3), 16GB of buffered ECC RAM, and an i3-4360/3.70GHz CPU. Once > the initial backup is sent, there will be little overall change in the data > (less than 0.5-1% change daily). Most of the change will be in the folders > that deal with my business documents, unless a family video or two gets > uploaded by my wife. > > Does this information help? Yes. So, the fact that it takes a while to chew through that data is normal, but it's probably not all that important anyway since you probably wouldn't be able to upload more than 50 Mbps anyway. ;-) Creating subsequent archives would be faster, since tarsnap can look at file metadata and figure out that the file hasn't changed -- thus avoiding the need to read the file and process it again. Does that make sense? -- Colin Percival Security Officer Emeritus, FreeBSD | The power to serve Founder, Tarsnap | www.tarsnap.com | Online backups for the truly paranoid