Disaster Recovery with SBS Backup Tools
| The sole reason for a backup's existence is for the data stored on it to be successfully restored at some unforeseen point in time. It is not there to ward off disaster as if by magic. All backups, be they recent, or archival, should be regularly tested by restoring some or all of the data and ensuring that that data is both readable and usable (can be copied, changed, and so on). At least a partial restore by selecting random files should be done on a monthly basis. This can be done by the backup operators themselves. Additionally, on a more infrequent and random schedule, a complete "bare metal" system restore to a test server should be done by the responsible IT. A surprise visit to the client to pick up the backup is a good way to check the client's readiness and the efficiency of the backup plan. The test restore also ensures that both the backup operator and the IT are familiar with the process and ready for any contingencies that may present themselves in the event of a real disaster. On a day-to-day basis, the existence of the shadow copies and of the Exchange deleted item retention, means that users themselves can restore missing data without recourse to the media and backup and all the time and effort entailed in restoring from that. Restoring a BackupTo do a full system restore of SBS 2003, you must first install the core operating system on the server. In SBS, this means that when you get to the part of installation where the integrated setup starts (see Chapter 4, "Installing SBS 2003 SP1 on a New Server," for more information), you cancel the setup process so that you only have the Windows Server 2003 portion installed but none of the SBS-specific components. Next, you need to install any operating system service packs that were installed when the backup was taken. For example, if you had installed Windows Server 2003 SP1 as part of SBS 2003 SP1 and you reinstall the operating system with pre-SP1 installation media, you will need to install Windows Server 2003 SP1 before you can continue with the restore. Next, you will run NTBackup to restore the data from backup onto the server. Should the hardware be substantially different from the original, the system may need several reboots into Directory Services Recovery mode for the appropriate drivers to be loaded. This is why it is imperative that the disaster plan be practiced on both similar and dissimilar hardware. Microsoft recommends that the following items on the new server should be as close as possible a hardware and version match to the original:
You can use an IDE controller instead of a SCSI controller. Also, the hard drive on the new computer should have the same number of volumes or partitions, and they should be at least as large as the originals. If the network cards have changes, you need to rerun the Configure Exchange and Internet Connection Wizard (CEICW) so as to re-create the settings with the new MAC addresses of the network cards (NICs). You may also need to reregister the server and reinstall additional client licenses. By rights, disaster recovery deserves a chapter, if not a book of its own. There are a multitude of server configurations, and each one brings its own unique requirement to the recovery process. This is why it cannot be stressed strongly enough that the disaster plan for each business must include the testing of recovery to both identical and dissimilar hardware of that business's backup. It cannot be assumed that what worked for business A will work for business B when it comes to disaster recovery.
For more information about the backup process on SBS, review the article "How to Back Up and Restore Your Server" from the SmallBizServer site (http://www.smallbizserver.net/Default.aspx?tabid=144). |
EAN: 2147483647
Pages: 253