I have a MD3620f with two controllers (slot 0 and slot 1), configured with 3 virtual disks, where two of them have preferred path as slot 0 and the other (the third VD) uses the slot 1 as preferred owner. When I run the "Verify Disk Media" of "! QLogic Fast UTIL "in the VD that uses slot 1 everything is okay, but when I run the other two VD's which preferred owner is slot 0, returns me the following error:
SCSI Command Error
SCSI operation code: 2F
SCSI sense key: 06
SCSI additional sense code: 8B
SCSI additional sense code qualifier: 02
Can somebody help me?
So, it would appear that you're using the HBA to verify your media, and it's not going to be able to do that. Your HBA can see the SAN, but it can't see the drives. The SAN controls the drives. As far as I know, the only thing you'd use the HBA utility for is to possibly set 'timeout values'. Everything else should be done in MDSM.
In short, an unclear communication trying to verify media with that utility is to be expected. Is there a reason you're not using the MDSM for this?
I'm also using MDSM. I used "QLogic Fast Util" just for a test because I didn't get the virtual disks on. But now it's working.
The problem is, now always that I turn off/on the MD3620, it led status is blinking ambar and a fault is launched into MDSM, so the storage automatically change the preferrred owner from slot 0 to slot 1.
Why does it occur?
What are the errors showing in the 'Recovery Guru' in your MDSM when it comes up? Does it [VD Ownership] stay on slot 0 until the array is booted "off/on" and then fail over?
The blinking LED on the array simply tells you that there is a fault somewhere. It has no way of telling you what that fault is. This is why we depend on the MDSM to report those errors.
Failing over from one slot to another can be from many things.
Would you be able to attach a support bundle and let me look it over?