Intel RAID Controllers - Best Practices white paper

Intel® RAID Controllers Best Practices White Paper

Revision 1.0

 Partially Degraded – A virtual disk with a RAID level capable of sustaining more than

one member drive failure experiences a member failure but the virtual disk is not

degraded or offline.

 Degraded – A virtual disk that already has one or more member drive failures and

cannot sustain a subsequent drive failure.

 Offline – A virtual disk with one or more member drive failures that cause the data to not

be accessible.

8.3 Tips and Tricks

8.3.1 Setup Tips

 Check cables for proper connection.

 Verify that all the cable ends are properly seated and that none of the pins are bent.

 Verify that an approved cable is being used. Cables must be speed compatible and meet

signal integrity specifications.

 SATA cables are designed to connect directly from the RAID controller to the hard drive

or drive enclosure.

8.3.2 Debug Tips

 Improvements in RAID controller and hard drive communication and control are

frequently incorporated into updated versions of RAID controller and hard drive firmware.

It is generally recommended to review the release notes for firmware updates and apply

the updates as warranted.

 Review firmware updates for the server board and intelligent backplane and complete

updates as necessary.

 Grown defects may not indicate that the drive is failing. However, if the number of grown

defects is large or increasing, the drive may be in the process of failing. It is

recommended that you replace the drive.

 Bad block redirections may not indicate that a drive is failing. However, if the number of

redirections is large or the number is increasing, the drive may in the process of failing.

It is recommended that you replace the drive.

 Parity errors in a log may indicate a failing controller, failing drive, or a memory issue on

the RAID controller. Replace the controller and / or hard drive. Some RAID controllers

include a DIMM site; verify that the memory used is listed on the RAID controller’s tested

memory list. If errors persist, try changing the memory module and / or RAID controller.

 Write compare errors indicate a failing controller or failing drive. Replace the controller

and / or hard drive.

 Do not reinstall a drive that has failed.

9. Summary

Intel is committed to providing customers with a stable product that offers both high-

performance and high-reliability. In this document we have provided guidance on cache options,

using a RAID controller battery, using a UPS, and other important options that can enhance

data safety and controller performance.