Link to home
Start Free TrialLog in
Avatar of changjia
changjia

asked on

Best Practices - Full Restore of SBS 2003 onto a new array

I'm looking for the best practices way to back up and restore SBS 2003 to a new hard drive array on the same server it's currently running on.  I've had success in the past using the Windows backup to take a full system backup, installing SBS on the new array, and then doing a restore from the full backup.  This last one we did failed and the machine bluescreened after the restore.  We ran out of time and couldn't run a repair on the restored OS (would have probably fixed the problem) but we had to roll back.  I found two things that I didn't do that may have caused it.  
1. We didn't make the C & D partitions the exact same size as they were on the old array, and
2. We didn't bring the newly installed OS up to the same Service Pack as what was running on the old array before I ran the restore.
We would much prefer to do this as a bare-metal restore but the only product I know that will do that is Symantec Backup Exec - System Recovery and we can't spend a $1K just for this restore.  Microsoft says that the blue screen is caused by the hardware change and 97% of the time can be be fixed by running a repair.

I need to know how you guys and most large enterprises are doing this to establish what the best practice should be for our organization.  Your thoughts and recommendations are greatly appreciated.
ASKER CERTIFIED SOLUTION
Avatar of Glenn Abelson
Glenn Abelson
Flag of United States of America image

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
Re:
1: Size of partitions (as long as they are large enough) does not matter. If you had Drives C and D it will be easier if you recreate them on the new array.
2: SP levels have to be the same otherwise you are doomed.
3: Make sure USB drives are disconnected while doing restore.
Hope that helps,
Olaf
Restores can be very messy, and every situation can warrant a different approach.  For example, our company has terrabytes of data, and can take up to 48 hours to restore all of the data we host.  So bare metal restore is not an option (or we're in big trouble).  We keep a warm standby that is within a week or two ready to go.  On our production machines, we run mirrored drives and clusters.  So if a computer completely dies, an entire other server is ready to take over and continue on.

Let's look at your situation, which doesn't sound too bad.  If you don't want to make the investment in Veritas just yet, another option might be to take a fresh hard drive that is the same size as the existing one in the SBS server, then install that hard drive into the SBS server, taking out the existing drives.  Install the same operating system, put on the service pack, and then remove the hard drive and stick it in a closet (or a safe place).

Then, on the day that you need to do a full restore, pull that hard drive out of the closet, boot up, run the restore, and you're done.

If a restore method is imperative that it work, then be proactive about it and try the method before you buy into it.  Even with Veritas, try the bare metal restore before you rely on it.  The reason for that is even Veritas bare metal restore fails half of the time, unless you did it just right.  Or didn't have a certain service pack installed.  Or had a patch installed.  etc...  It's better to be proactive and find that you have a restore problem, and get that problem resolved, before you need to actually do a restore for real.
Avatar of changjia
changjia

ASKER

glen_1984,

The tech republic article was very helpful.  As a matter of fact we brought in the author as a Subject Matter Expert to consult for a few hours on our server upgrade.

Thanks.
Glad it worked.
I could have used him last week.