HP A7143A RAID160 SA Controller Support Guide, February 2007

RAID160 SA Controller Overview

Board Components and Features

Chapter 2

NOTE An online spare does not become active and start rebuilding when the imminent failure

alert is sent, because the degraded disk has not actually failed yet and is still online. The

online spare is activated only after a disk in the array has failed.

• Drive failure alert features cause an alert message to be sent to Event Monitoring Services (EMS)

when physical disk or logical drive failure occurs.

• Interim data recovery occurs if a disk fails in fault-tolerant configurations (RAID level 1 or higher).

If a physical disk fails in RAID 1, RAID 1+0, RAID 5, or ADG, the system will still process I/O requests,

but at a reduced performance level. Replace the failed physical disk as soon as possible to restore

performance and full fault tolerance for the logical drive it belongs to.

The risk of continuing operations without replacing a failed physical disk varies depending on the RAID

level that has been configured:

— RAID 1: If RAID 1 is configured, the result will be a single mirrored pair of disks. If one physical disk

fails, the remaining disk in the mirrored pair can still provide all of the data.

— RAID 1+0: A RAID 1+0 configuration will have a minimum of 4 physical disks, and the total number

of physical disks will be divisible by 2 (to support mirrored pairs). In RAID 1+0, if a physical disk

fails, the remaining disk in any mirrored pair will still provide all of the data that was on the failed

disk. In fact, several physical disks in an array can fail without incurring data loss, as long as no two

failed physical disks belong to the same mirrored pair.

— RAID 5: If a physical disk fails in a RAID 5 configuration, data is recovered via a parity formula and

is typically written to an on-line spare physical disk. If a second physical disk fails before the data

from the initial physical disk failure has been rebuilt on the on-line spare disk, the logical drive will

fail and data will be lost.

— ADG: Similar to RAID 5, ADG also relies on a parity scheme to rebuild data if a physical disk fails.

However, in an ADG configuration the parity data is duplicated on two different physical disks. As a

result, ADG can support the failure of two physical disks without data loss.

For a more detailed description of the RAID levels supported by the HP A7143A RAID160 SA controller

see Chapter 1, “RAID Technology Overview,” on page 11.

For detailed information on the probability of logical drive failure, see Appendix A, “Probability of Logical

Drive Failure,” on page 93.

• Recovery ROM is a redundancy feature that ensures continuous system availability by providing a

backup ROM. This feature protects against corruption of a ROM image (caused, for example, by power

fluctuation during ROM upgrade). If corruption occurs, the server automatically restarts using the

remaining good copy of the ROM image. When you upgrade the ROM, the inactive image (the one not

being used by the system) is upgraded. There is not normally any noticeable difference in operation.

When you use Recovery ROM for the first time, however, both ROM images are upgraded, causing a boot

delay of about 60 seconds.