Q.)  Is it possible to BREAK a running mirror on a ProLiant server connected to a Smart Array controller, running either LINUX or Windows? 


A.)  There is a whitepaper here:




When using a configuration with RAID 1(+0) volumes attached to Smart Array controllers, it may be necessary to break a mirror and rebuild the drives. For example:


? When a drive in a RAID 1(+0) volume fails, the controller essentially breaks the mirror. An administrator subsequently can regain fault tolerance and resynchronize the mirror by inserting a replacement drive and allowing the controller to rebuild the drive automatically.


? As a precaution against a future failure situation, an administrator can break a mirror by removing half of the mirrored drives (which are saved as the system backup), replace them with other drives, and allow the controller to automatically rebuild the new drives to restore fault tolerance. (Reverting to a previous backup is more complex and is explained later in this paper.)


? When increased capacity is required, an administrator can replace half of the drives in a RAID 1 or RAID 1+0 array with larger capacity drives and rebuild the new drives. Afterwards, the administrator can replace the other half of the drives and allow them to rebuild in the other direction. Once all drives in the array are of larger capacity, the administrator can use the Array Configuration Utility (ACU) to increase the logical drive size or to configure additional volumes.


? When troubleshooting drives or testing software configurations, an administrator can use half the mirror drives as a backup copy and then revert to that backup copy if a software configuration (for example, an operating system patch) is unsuccessful.