The database format was chosen more to minimize storage requirements on the server than to deliver high performance or great robustness, I believe. Conceptually, the backup database is "backed up" by the clients that are backed up in it; if something happens
to the database you just back everything up again. I'm fine with this; in all my years as an IT pro, I have never restored farther back than "last proven good" backup. Even in the case where a patch corrupts an OS, you still just step back to the backup just
before that patch. And you generally know almost immediately if a patch causes problems. The only time you would have to go farther back is if backups have been failing, and if you test your disaster recovery plans regularly (you
do test regularly, don't you?) you'll know about that pretty quickly.
There has to be a path to migrate to Vail without losing all current data, right? I've not tried the beta, but this all sounds like another example of MS losing sight of the basics at the expense of something else. I never worked on WHS, but I did work there
for a long time. I know that shipping is a feature too, but...
Copying from one server to another is one option. Another is to back up your shares form your V1 server, then copy the contents of the backup to the Vail server (they're just files). But if you want to re-use the hardware you
will have to flatten the server. As I said, upgrading isn't possible in general when moving from a 32 bit to a 64 bit OS, and Vail is a complete rewrite of Windows Home Server. There's almost nothing except basic concepts that the two
have in common. Combine that with the fact that Windows Home Server is an OEM product, not shelfware, and you can see why upgrades aren't a consideration: Microsoft expects that the vast majority of users (who bought OEM servers from HP, Acer, etc.) will just
buy a new server.
I'm not on the WHS team, I just post a lot. :)
