I have a server on one of my networks that is running Exchange 2003. Yes, we will be upgrading to 2010 in about 6 months with new hardware and all, but for now I need to make sure this is stable and safe. So for now I wanted to ask advice for backing up and disaster recovery of this server. Here are the specifics:

This server is running Windows 2003 Server and Exchange 2003 with all updates and patches. The hardware is a Dell PowerEdge 1850 with no RAID card. The OS and data partitions are both on a single SCSI drive in the first slot 0. Right now the second drive is blank.

We are backing up the server using online NTBACKUP, and a hard disk Barracuda solution that uses a backup client to copy up the mail store to a local backup server and keep copies off site as well. We currently have no brick-level or offline backups in place.

In protecting this server my first thought is the drive. I can create a software mirror to the second drive, but I will need to convert the disk from basic to dynamic to do this. I am a little scared to do this, since I seem to remember doing this many years ago and having a problem (like the disk didn't boot after the conversion or something).

We just purchased an identical PowerEdge from overstock.com (believe it or not), so I have that to work with. And I also have several blank SCSI drives I can use for this.

So my question is what would you do to protect this server, knowing how it is set up and what I have available?

1 Answer 1


Mirrors are are for high-availability, not backup. NTBACKUP is the easiest bang for the buck for a single isolated WS03/Exch03 server.

  • +1 - RAID isn't backup. Having said that, I'd go ahead and mirror the disks using Windows software RAID and keep right on running NTBACKUP-based backups. Between that and the spare server you ought to be in good shape. As long as you're getting good, regular backups you're protected from corruption. You're limited to two disks in the PE1850 so there's no way you're going to be able to put the ESE transaction logs onto another physical spindle and still have RAID-1. I'd rather have RAID-1 than log/database separation anyway, personally. Mar 12, 2011 at 2:43
  • 1
    +1 raid is not backup. Raid is for increasing availability. However, ntbackup can be a bear to restore from if the file size grows too large. Make sure you test your backup's ability to restore! If you can, then bravo. If not, it's time to scramble. I've had luck with fpns.net/willy/msbackup.htm working on a 250 GB file...
    – tsykoduk
    Mar 12, 2011 at 6:42
  • Thanks for the comments. I know RAID is not backup - I was more asking what to do to protect in general. And specifically the question of converting disk 0 from basic to dynamic. Is there any realistic risk when I pull the trigger on that convert? If I can get that to work, then I will have a software mirror + 2 different backups to go to in case of corruption.
    – charnley
    Mar 12, 2011 at 15:29
  • Excellent point about testing the backup. Especially since you have the spare server: I'd say test the restore on that, do the dynamic conversion on it, and see how it goes. Maybe even switch over to that if you can afford some downtime. In theory converting to dynamic should have no visible effect (except for the features) but we all know how useless theory is. This KB article has all the steps for converting to a mirrored drive as well as links to other ones for managing/reparing etc: support.microsoft.com/kb/323432
    – Mark Sowul
    Mar 12, 2011 at 17:06
  • Just to let everyone know the conversion from basic to dynamic hosed the disk. I had to use a low-level hex editor to modify a sector to get the disk back to basic. Moving forward I will be using a third-party software solution to clone the drive for disaster recovery. Beware of converting from basic to dynamic!
    – charnley
    Mar 15, 2011 at 17:06

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .