On Mon, 25 Oct 2010, Derek Simkowiak wrote: > I have a client to needs to take "a lot" of data (many TB) and >archive it off-line. We are going to use 1.5TB disk drives as the >off-line storage medium... > > The trick is that we want two identical copies (one on-site, one >off-site). We could do it the old-school way, one disk at a time, but >this archival process will be an on-going need. So we want a way to >make the two copies "automatically", and without having to manually >process and catalog two disks every time. > > The best solution (I think) would be to use an eSATA disk case, >similar to an external USB or Firewire chassis, which has hot-swappable >bays, and a built-in RAID-1 controller. Then we just pop two disks in, >make the archive copy (and catalog it), and pull the disks out. Then we >hand one of the two disks to the off-site manager and be done.
Why use RAID, especially hardware RAID? If you wish to be able to restore from one of the disks, then I don't see what RAID gets you. As for writing two "identical" disks, is it essential that they be identical? And how fancy does your catalog need to be? I'd say use rsync (or whatever) and just create the two disks independently of each other. Yes, in theory, you'll be reading each file twice from the source disks, but with good file buffering this shouldn't be an issue (and the writes will be the controlling rate anyway). So: Plug in the destination disks, run rsync to one and (perhaps concurrently) to the other, remove disks ... Go with simple, life's complicated enough as it is ... -- Mike Schuh, Seattle USA
