Error on drive SCSI ID 6? - There is no SCSI ID 6 in server!

[James Mulgrew]

Hi folks -

Okay, got a weird one here. Got a DL380G4 server with 6 drives in it and a 6i smart controller with two internal "ports". When this server was built some time ago, the engineer elected to put the OS on a mirror on the drives on port 1, SCSI ID 2 and 3. They then created a RAID 5 array using port 1, SCSI ID 4 and 5, as well as Port 2, scsi ID 0 and 1. I'm sure they had a good reason for that. I've included a snapshot of the array configuration (physical view) so it can be referenced.

One of our engineers noticed recently that the server in question was reporting a 1787 error on POST, reporting that the drive on SCSI Port 1, SCSI ID 6 was operating in interim recovery mode. We don't have a SCSI ID 6 in the server! (grin) The scsi drive that is inserted in port 1, scsi ID 2, is not lighting up and it's been popped in and out, and replaced, a couple of times it is night lighting up and that is because the server sees it as available and not assigned to an array. I've attached a screenshot of the POST error.

HP's IML is not showing the post error but the Windows system event viewer is. The server was built at one point with the OS being mirrored on SCSI IDs 2 and 3 but now it thinks the mirror is with scsi ID 3 and 6 which does not exist. I've taken all 6 drives out of the server and popped in 6 new ones, configured them identical as to how the server was configured to begin with. I get no POST errors. I pop the original 6 drives back in to their positions and get the error again. Bizarre, I've not seen this before.

Is there any way to correct the issue without losing the OS? Thanks.

[EAJohnson]

Looks like there WAS a SCSI ID6 at some point and it failed... this is a strange one.

To fix it, I would suggest migrating Array A down to a RAID 0, then add SCSI ID2 back to it and migrate it up to a RAID 1.

I've never had to do that, so I can't verify it works. But right now you're down to one drive for your OS. If you keep inserting and removing them, you risk that drive encoutering a physical error and then you've lost your OS anyway.

Make sure you have a backup of ARRAY A before trying anything!

Good luck.

[James ~ Happy Dude]

Hello James,
http://h18007.www1.hp.com/support/files/server/us/download/15125.html
Please post the ADU report.

Regards.

[Brian_Murdoch]

James,

I don't know if it will help but firmware version 2.80A for the 6i apparently has a fix for unexpected SCSI ID changes.

Quote from it's history file.....
"Fixed Array behavior to prevent unwanted device migrations by detecting conditions, where a SCSI disk drive unexpectedly changes its SCSI ID"

http://h18023.www1.hp.com/support/files/server/us/download/27980.html

Perhaps this upgrade may help sort out the correct SCSI ID's. I suspect that SCSI ID 2 is also the phantom SCSI ID 6.

I hope this helps,

Brian.

[Víctor Cespón]

I have seen this a couple of times, where the controller reports an inexistent drive as failed, an the RAID does not rebuild because it's waiting for that drive to be replaced.
This is due to a one bit change in the RIS, the RAID included SCSI ID 2, and a single bit change transforms the 2 into a 6. The controller now is reporting ID 6 as failed and expecting it to be replaced. Of course, there's no SCSI 6 on a DL380, and no easy solution for this except deleting an regenerating the RAID.

[James Mulgrew]

I've tried running the ADU with no progress. Let it run yesterday for 8 hours and it never completed.

Updated the firmware as well but the problem still exists.

Unfortunately I can't make any changes to the array itself (migrate, expand, etc) while it is still in an error state.

I am afraid the only way this is going to be resolved is to wipe it out and rebuild.

- Jim

[James ~ Happy Dude]

Jim, Is there a Cache Battery Connected to the Controller ?

Regards.

[James Mulgrew]

No, there is not. I had thought of the possibility of removing the battery as well. (grin)

[James ~ Happy Dude]

I thought so.

Try booting from SmartStart.. Goto ACU;
Change the rebuild priority.. save n exit... reboot.
boot from SS CD Again... Change rebuild priority back to normal... Reboot.

Try it.. I would be surprised myself.. if this works.

Regards.

[James ~ Happy Dude]

Did u fix it ?