Then you're probably best looking at a replication technology like Veritas Replicate Exec, Microsoft Data Protect Manager or NSI Software Double Take. The initial synch will take days, but after that it's much lighter: RE and double take replicate blocks in real time, DPM is hourly for example.
Some support system state, some don't, so you'll want to use NTbackup to take the system state once in a while.
I'd probably synch the data areas, but not the OS. Do a system state periodically to another replicated folder. If you really need the rest of the OS (like program files), use NTbackup for this too. Since changes to the OS are rare, schedule this NTBAckup for Friday night so you have all weekend for it to replicate over to the other side.
we use double-take for our exchange server - we can fail-over to our D.R. site and have our 1400 odd users accessing their mail within 20 mins - and most of that delay is verifying DNS etc is updated throughout the domain. It's brilliant, but expensive if you only want data replication (ie no need for name failover and no sql or exchange to worry about).
I have RE in trial at the moment to centralise my back-ups (like I said, DT was too expensive for this unfortunately). Most of the servers took around 3 or 4 days to complete the initial synch (throttled to 75% of overall bandwidth - bandwidth from 128kbps to 1mbps, data volumes of 5 - 35GB) and thereafter use virtually no bandwidth. Only disappointment is the lack of logging and reporting that's available.
DPM is much cheaper and I intend rolling that out for trial next week.
Alternatively, don't underestimate the value and bandwidth available via sneakernet! Simply backup to a tape device, and ship the tape to the 2nd site where you can restore - restoring daily proves your back-up is working well and keeps the data upto date
|
