aix 5.1 disk errors on eva5000

[lorance]

Has anyone seen disk errors that are associated with the ccls on the EVA5000 that are reported in the errpt on an aix server. These errors just started about 2 weeks ago and occur every 4 hours. There ccls are not in a volume group but something seems to be accessing them. Any ideas would be appreciated.

[Uwe Zessin]

Do you know if there was any software change on the servers or the EVA? Some diagostic software?

The CCL (command console LUN) is not a disk device and, I think, never was. It is a communications channel for the Storage Management Appliance and allows the servers to find their LUNs (SCSI report luns).

[lorance]

To my knowledge there was no updates to the EVA. We are still investigating any changes to the server. I have never run across this type of disk error that points to the hdisks on the eva ccls. If the disks are not in a volume group, how can the o/s access them?

[Mike Naime]

Lorance:

What exactly is the text of the error message?


My experience with AIX servers and a Compaq/HP SAN is that the CCL reports any maintenance activities within the storage controller out to the AIX systems that are attached to that storage system.

HSG80 Experience: If I have a drive failure in a storageset that is attached to another server, that gets reported out. If I replace a failed disk drive, that gets reported out. If I insert or remove disk drives, that gets reported out.

Note: The same ambiguous "Check your storage system" message gets reported for all of these activities.


EVA Experience: Drive failures get reported out.


Mike Naime

[Uwe Zessin]

lorance,

the CCL is a SCSI LUN, but not a disk device. Different LUNs can have different device types (disk, tape, changer, scanner, CD-ROM, array controller, ...).


Mike,
that's interesting...

On the Host Properties, there is a selection 'Direct Eventing'. According to the online help (CV-EVA V3.1 / VCS V3.010) this means that the EVA can send events directly to the host. However, it can only be enabled for Tru64 Unix and Windows hosts.

Is any of you running an SWCC agent on the AIX hosts, too? (just an idea...)

[lorance]

The output is from aix errpt.
No there is no swcc on this server.


Resource Name: hdisk5
Resource Class: disk
Resource Type: HSV110
Location: 11-08-00-0,0
VPD:
Manufacturer................COMPAQ
Machine Type and Model......HSV110 (C)COMPAQ
Part Number.................
ROS Level and ID............33303130
Serial Number...............4889INY0
EC Level....................
FRU Number..................J0P
Device Specific.(Z0)........0C000212F7000032
Device Specific.(Z1)........4F4889IM804A
Device Specific.(Z2)........
Device Specific.(Z3)........
Device Specific.(Z4)........
Device Specific.(Z5)........
Device Specific.(Z6)........

Description
DISK OPERATION ERROR

Probable Causes
DASD DEVICE

Failure Causes
DISK DRIVE
DISK DRIVE ELECTRONICS

Recommended Actions
PERFORM PROBLEM DETERMINATION PROCEDURES

Detail Data
SENSE DATA
0A00 0000 5600 0000 0000 0000 0000 0000 0102 0000 7000 0500 0000 0007 0000 0000
2000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000 0000
---------------------------------------------------------------------------
LABEL: DISK_ERR2
IDENTIFIER: A668F553

[Toby Foote]

Yes, we are having this same problem with AIX and EVA. And we've determined that the problem is our Best1 BMC software accessing these disks (every 4 hours, just like you are experiencing), even though they aren't in LVM. I suspect you have something similar.

I'm still working on resolving this issue, as we are continuing to get Tivoli alerts on this. Best1 BMC is suggesting we turn off DISK_ERR2 logging in the errpt (ha!).

If this is what's happening to you, please contact me. I'd like to talk to someone else having this problem, as IBM, HP, and BMC don't seem to know how to fix this problem. Each one is telling me "this is just how it works and you have to deal with it". I don't especially like that answer.

Talk soon,
Toby Foote
216-689-9340

[Uwe Zessin]

If the error is about EVA's LUN 0, then BMC should check the device type before they do inquiries. LUN 0 is not a disk device.

[Toby Foote]

Uwe,

BMC touches all hdisks, and that includes the CCLs that are the fibre paths to the SAN LUNs. That's when the EVA freaks out and reports something to the error log.

Toby