I have a SCSI RAID controller and the drives keep dropping offline
I have a SCSI RAID controller and the drives keep coming up marked dead/off-line. I can return the drives on-line again, but why is this happening and what can I do to fix it?
Drive randomly marked dead (off line) by SCSI controller.
There are many reasons why a SCSI controller or operating system marks one or more drives off line. Here is a list of common issues, likely to cause a drive to be remarked off line–dead.
It’s important to make sure the cabling, termination, and drive enclosure hardware is suitable based on SCSI specification–or as required by your configuration. Use the best quality SCSI data cable. Make sure it is certified for either Ultra SCSI or Ultra 2 or Ultra 3(Ultra 160) or Ultra 320 SCSI. Most SCSI controller vendors support a line of cables and terminators for their product.
If there is a mix of LVD (low voltage differential) and non-LVD devices on any one SCSI channel? Then Ultra SCSI (single-ended narrow) specifications apply–cables not longer than 1.5 meters, and not more than four devices supported. Both LVD cables and termination devices are available from: Granite Digital at 510-471-6267, and TMC at 415-454-5425. A distributor or computer reseller may sell LVD SCSI data cables and terminators.
Not enough SCSI Termination Power can be at fault. As a precaution, be sure the last two physical devices (hard drives) on each SCSI channel, provide Term. Power to the SCSI bus. For SCSI bus Termination Enable, we highly recommend you use either the disk array enclosure (if supported) or a system terminator at cable end. If you use LVD drives, you must use an LVD terminator. If you are using SCSI single ended devices, or mixing single-ended and LVD you must use an active terminator at cable end.
Sometimes the disk array enclosure, the drive slide trays may be defective, or there could be a problem with the SCA back-plane logic board. Try the drive in a different system–does the failure follow to another ‘known good’ platform?
The system power supply can be at fault. Under or over current will cause the drive to spin down in the name of data integrity. You will want to make sure a high quality power supply is used, and that it will meet the power needs of your system.
Software drivers and BIOS on the SCSI RAID controller can affect timing and communications. You should visit all system component and operating system manufacturer’s Web sites, to confirm that all drivers and firmware is up to date.