PE6600 RAID problems

[biglebowski]

I have 14 6600 with powervaults connected via scsi. Since the servers have gone live (6 months ago)4 of the servers have had disk failures corrupting the oracle database on one of them and causing another to BSOD. All the hard disks that fail are onboard not the powervault. They are running RAID1 which doesn't appear to be working. Has anyone seen this before or have any ideas why it may be happening?

That rug really tied the room together, did it not?

 

[technome]

A little info would be helpful...

Please supply the service tag, as this will make answering some of the following unecessary.

Did Dell setup the disk mirrors or was this done in house?

Are the disks off a raid adapter or SCSI interfaces?
Please give raid adapter scsi interface models.

If in scsi how are the disks attached, to the onboard scsi interface(s) or addin scsi adapters? Are the mirrored drives on separate scsi adapters/interfaces ?

Are any other devices attached to the scsi buses the disks are on?

Are the supplied cables high quaility?

Machine bios, and scsi adapter or raid card firmware and drivers at the newest level.

........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm

 

[biglebowski]

Ok...
I can't post the service tag as it's a DOD server but heres the rest:
DELL didn't setup the RAID it was configured by a vendor.
The array disks are connected to a PERC 3/DC controller.
It is listed as a PERC LD 0 PERCRAID SCSI disk device
There are only hard disks connected to the PERCRAID device.
The f/w vers are:
PERC 3/DC 196T
Baseborad 1.78
Prim backplane 1.01

That rug really tied the room together, did it not?

 

[technome]

I would do this only on the machines affected.....

I would update the machine bios to the latest

Re-seat all the scsi connectors,the raid adapter.

If you have Seagate drives, 10 or 15K, firmware >006, there is a firmware update from Seagate, due to corruption problem.

Your raid firmware is not the latest, not sure of the backplane.

Run consistency checks scheduled weekly, off hours if possible, on all the servers.

run chkdsk on all the servers

BSOD, probably unrelated to the raid drivers, unless the firmware and driver combination are not closely dated. I would update EVERY device driver in device manager, including the chipsets.

........................................
Chernobyl disaster..a must see pictorial

http://www.kiddofspeed.com/default.htm