ML350 G4 Hard Disk problem

[chris_p]

Hi,

I have an ML350 G4 with a smart array 641 controller.

The server origionally had 2 disks installed in a mirrored configuration, one of these disk then failed. The server was also low on available space so I purchased 3 new disks, one to replace the faulty disk and 2 to create a new mirrored set.

All 3 disks are the same and certified for use with this server and controller and are non hot-plug.

I used the jumpers to assign the SCSI ID's

I installed 2 new disks and created a mirrored set, moved across exchange data etc. No problem

Problem:

I then wanted to replace the faulty disk (SCSI ID 1), after instaling the disk and powering up the server I got 2 POST error's, 1775 and 1785, the server would not boot. With this disk connected OCRA does not detect any disks, the ACU detects all 4 disks but shows them as unallocated.

I then updated the firmware for the 641 controller and the BIOS, same problem. Removing the disk then allows the server to boot normally, the ACU shows the new mirrored set and the failed mirror( Interim recovery mode).

I then took this new disk and plugged it into a different HP server, the disk was recongised in OCRA and ACU as SCSI ID 1, so the disk is OK.

I have checked the SCSI cable for bad PINS, I can alos put the failed drive back in the server and it boots without POST error.

I then updated the firmware on the disk as instructed by HP and installed it agian, same POST errors.

Also ran the offline diagnostics which did not produce any errors.

If anyone has any ideas it would be appreciated.


Thanks,
Chris

[wobbe]

If I'm not mistaken post 1775 has to do with cabling issues and post 1785 is something about the array configuration.
Did you use the same scsi id twice?

I would, switch off the server and take out all disks. Remove any configuration that is on the raid controller. (not sure if this is neccesary)
Only insert the one surviving disk of the original configuration. Make sure it uses the same scsi id and same sici bus as before.
If it works the controller will read the original raid configuration from this disk and you should be able to start. You'll get an error because of the failed/missing disk and you have to hit f1 or f2 to continue booting. (think f2)
If it doesn't find any raid configuration then you are f....d.

[chris_p]

Thanks for your post. I was onsite yesterday to try and resolve this issue.

When I have the 3rd disk connected I see the same SCSI ID twice. Also no logical drives detected by the smart array.

As a test I removed one of the new working disks (SCSI ID 8), this is mirrored with SCSI ID 4, booted and then it showed that both mirrors were degraded, I then took the "Problem disk" and assigned it SCSI ID 8 and connected it in the same position as the previous working disk, since this disk is basically in a factory default state I was expecting it to rebuild the data from SCSI 4, but no, the same 2 post errors again. So this ruled out a cable problem

As you say the the 1775 error advised to check cable and termination and the 1785 advises that the smart array card is not responding (EMU not responding I think).
I have reseated the array card and checked the cable and pins for anything obvious.

I will try what you have suggested and post back the outcome.