Recovering A Stripe Set With Parity

Recovering a stripe set with parity is not going to cause you to have heart failure, nor will it require using any disk tools, boot disks, or other utilities. In fact, if you have sufficient free space on another drive, you might not even need to open up your computer. If you do not have sufficient free space, you will need to stop the computer, replace the failed drive, then follow the steps presented in this section.

1.  Launch Disk Administrator.

2.  Select the recoverable stripe set with parity.

3.  Select an area of equal or greater space on an unpartitioned drive.

4.  Choose Regenerate from the Fault Tolerance menu.

5.  Close Disk Administrator, and restart your computer.

When your computer restarts, use the Volumes view of Disk Administrator to check the status of the stripe set with parity. The status should be initializing, which indicates that the missing data is being regenerated. If you get an error message that states the drive could not be locked for exclusive use, it means that there are open files on the stripe set with parity. This might be caused by having a paging file on the stripe set with parity, a service such as SQL Server or Exchange Server with database files on the stripe set with parity, or users accessing files on the stripe set with parity. If the cause is the paging file, you will have to move the paging file to another drive. This is a good idea anyway. Placing the paging file on a stripe set with parity degrades the performance of the operating system as well as the stripe set with parity. Use a stripe set (no parity) for the paging file for best performance. If the open files are caused by a service, you will need to shut down the service until the stripe set with parity has been rebuilt. If the open files are caused by user connections, you will need to kick the users off the system, then shut down the server service. You can do this gracefully by first broadcasting a message warning of the impending loss of service, pausing the server service to prevent new users from connecting, broadcasting another message with an imminent shutdown warning, then stopping the server service.

Summary

In this chapter, you learned quite a bit about how redundant disk and fault tolerant systems operate under Windows NT Server. You also learned how to create mirror sets, duplexed mirror sets, and stripe sets with parity using the Disk Administrator. Should an error occur during your tenure of office, you should be able to rebuild these redundant or fault tolerant systems. There are a couple of key items I would like to remind you of, however, including:

•  A mirror set provides a high level of fault tolerance for your system partition.

•  A mirror set can provide twice the read throughput, and a duplexed mirror set can provide almost four times the read throughput, when compared to a single disk controller/single drive combination. As such, a duplexed mirror set provides an excellent means of improving query response for an SQL Server database as well as providing additional fault tolerance.

•  A stripe set with parity provides improved read/write response in relation to the number of drives installed on the disk controller. Multiple disk controllers can provide even higher throughput. A stripe set with parity also wastes less storage space (the more drives—the less wasted space).

•  A stripe set, while not discussed in detail in this chapter, provides even higher throughput than a stripe set with parity, but a stripe set has no fault tolerance.