ML350G3, SA641, drive changed SCSI ID

[Sarandi]

Hi:

I have a problem with ML350G3 / Smart Array 641. Six drives attached (the whole cage). IDs 0 and 1 are mirror with boot Windows 2003 server. IDs 2, 3, 4, 5: RAID 5 with data. After a reboot, drive with ID 4 appears with red light. But raid status is OK, and ACU, ADU and Survey show now a new drive ID 6 and none drive with ID 4. Physically there is no drive ID 6. Events show that drives has changed position. I think about old FW, but where: in SA641? in drives? both? Another suggestion? SA641 FW is v2.34.

Regards,
Sarandi

[James ~ Happy Dude]

Hello Sarandi,

The latest Firmware for SA 641 is 2.76
http://h18023.www1.hp.com/support/files/server/us/download/26775.html

I would recommend downloading Firmware maintenance CD 7.90 & installing it.

http://h18023.www1.hp.com/support/files/server/us/download/27783.html

This will ensure ALL the firmware revision on your server is up to date.

Regards,
James.

[e4services]

Firmwares would not cause this physical problem. THe IDs should be physically, as you stated, fixed to the slot and should not float. AND ID 6 is not even used for the drive slots, specially out of order, so you got to suspect the backplane has a problem. I'd replace it.

[James ~ Happy Dude]

Hello,

The reason I suggested updating the firmware is, I had encountered a similar situation; it was an MSA1500; OS was RHEL & the Minor Number kept changing on every reboot;

Updating the Firmware then fixed it then;

:)

Regards,
James.

[Sarandi]

Thanks to all:

Last night I updated all FW and PSP both 7.90. Problem is the same: ACU shows both volumes good, but I still have a red light in one drive. After FW update, initial message from SA 641 changed and is more clear:
"1734 - slot1 Drive Array Disk Drive(s) Responding to incorrect SCSI ID. Backup your data, replace faulty drive, reconfigure array and restore SCSI ID 6".
Note the inexistent SCSI 6 to replace!. Now need press F1 to continue boot.

Next weekend I will go with a couple of drives, a backplane (if I get one) to take server off production, try each backplane position, reconfigure RAID, restore data, etc. I suspect backplane failure, too.