unable to use a disk with LVM

[Sameer_Nirmal]

Hi Guys,

I am facing a peculier problem with the system having L2000-44 connected to HP Disksystem 2110 ( A5675AD ) having 4 disks each of 36GB capacity.Recently we had a disk failure ( ID -2 ) in the 2110 box. Hence we got a new disk
(36GB LVD-SE 10K 320 Ultra-Wide SCSI ) , but it is not getting configured with LVM. If I do pvcreate on the disk, it says unable to write LVM record on the DISK, I/O error. I can see in the syslog that there is huge error being generated saying Resetting SCSI, Unhandled interrupt for that new disk ( c4t2d0 ) It keep on giving this error for couple of time before giving I/O error. Can anyone put more light on this issue and tell me what following error indicates?

lsp->state: 5
lbp->owner: 0000000000000000
scratch_lsp: 0000000042e0de00
script dump [000000004402d000]
0900dc00 00038000 e25c0004 f97ef7f4
80080000 f97ef090 80080000 f97ef090
SCSI: Resetting SCSI -- lbolt:42248, bus: 4
SCSI: Reset detected -- lbot: 42228, bus: 4
SCSI: Unhandled Interrupt -- lbolt:42749 , dev: 1f04200
lbp-> state: 2060
lbp-> offset: ffffffff
lbp->uPhysScript: f97ef000
ISTAT: 0a , SIST0: 18, SIST1:00 DSTAT: 80, DSPS: 000a8000

The 2110 is connected through SCSI-LVD controller ( 160 ) and is cascaded to a HP DAT drive and terminated. I removed LVM configuration from the OS pertaining to the vg of that disk ( we have one vg for each pv )and ran vgscan by disconnecting that disk from the slot to have fresh LVM configuration.But is gives that huge message!!!!!

The other 3 disk are of same type and configured with LVM and working normal...
Is it issue with the SCSI slot no.2 in the 2110?? How do i make out the cause from the above error?????...

Waiting for quick replies....

P.S. this thread ahs been moevd from Disk to HP-UX > LVM and VxVM. - HP Forum Moderator

[Alex Lavrov.]

You have a SCSI error on the disk.
The only way to resolve it it's to call HP support so they'll send a technitian over you and replace the disk.

SCSI: Unhandled Interrupt -- lbolt:42749 , dev: 1f04200

In this line you can see what device is this. Look at the live "dev:1f04200"

1f -> 31 - device driver address, if you'll do: lsdev | grep 31, you can se what kind of device is it (sdisk, stape, etc)

The other digits are c,t and d values of special file.

As I said, there is not much that you can do. Call HP.

[Norman_21]

Run this command and CTRL+C after 15 seconds, see if the disk will hang:
#dd if=/dev/dsk/xxxxx of=/dev/null

I agree, the following indicates a bad disk.
SCSI: Unhandled Interrupt -- lbolt:42749 , dev: 1f04200

I'd suggest that you install ISEE, it'll help log support calls with HP remotely.


Hope this help

[Robert-Jan Goossens_1]

Hi Sameer,

Agree you should call HP, big change this a DOA disk.

There is a warning doc in the ITRC for CU using Disksystem 2110 in combination with U320 HBA such as the A7173A, but since you use a SCSI-LVD controller ( 160 ) I don't think it will count for you (link is below)

Document description: Advisory: DS21XX and DS2300 Disk Systems Do Not Support U320 Speeds - Importance: CRITICAL, This Advisory addresses a potential data integrity issue.
Document id: OD041111_CW01

http://www4.itrc.hp.com/service/cki/docDisplay.do?docLocale=en_US&docId=200000076966737

Best regards,
Robert-Jan

[Mohanasundaram_1]

Hi Sameer,

before calling HP you can check few things yourself.

1) Remove the DAT drive from the SCSI bus and terminate the DS2100. Check if the problem comes. ( Plan to connect DAT drive seperately).

2) Check SCSI cable/terminator are properly fixed. Look for any kinks/bent pins in them.

3) if so, replace them and check.

4) As pointed out already, there is a big chance that the disk is Dead-On-Arrival. get another disk and test.

5) Ensure you got all the latest LVM patches.

Please revert with the result.

With regards,
Mohan.

[Sameer_Nirmal]

Thanks you all for your quick replies, appretiated...

I had done already my homework to check cables, pins ,separating DAT drive etc.. but the trouble continued.

One thing i noticed during my investigation about the disk. The disk is being identified as HP36.4MAP3367NC which firmware HPC6. All other disks are HP36.4ST3367LC and firmware HP00 . Does it make any difference??? I guess it shouldn't.....

I am gonna try with new disk and let's see what happenes.....

How much probability go u see of the slot being making trouble????

I referred the DS2110 doc , which says that
U320 drives don't operaye reliably at U320 speeds and are not supported at those speed.
That means even if I have U160 controller and U320 Disk, the disk should work at U160 speed without any trouble .

[Sameer_Nirmal]

Today I tested the original drive which was decleared faulty by previous SA in another working slot ( ID 4 ) . I just ran dd and diskinfo on that drive, it worked normal!!!!!
I am sure that this drive can be used with LVM. I tested my new drive ( A6571A ) as well and it worked too....

Now it's pretty sure that the slot ( ID 2 ) is giving trouble.. I would be inspecting the backplane and BCC tomorrow...
I doubt about if one can order chassis, controller separatly???

I will keep on posting the results..

[Bill Hassell]

You can order just about anything from HP. The (not so) secret location is: http://partsurfer.hp.com

[Sameer_Nirmal]

I have confirmed that the culprit is the slot......
Now i need to change the controller card of DS 2110 . I noted the P/N of the board. But i didn't get this part at HP parts sites.
the P/N is A5675-60001.
If i go the usual style of HP which follows the system board p/n for any system i.e
(system product number-60001) , then this is correct..... Any comment???

[Alex Lavrov.]

I think the best idea will be to call HP Support and talk with them. You'll get the best answers.