¿scsi error?

[Carlos_172]

Hi there,
Seems to be a problem whith some of our disks. Very often there is a proccess very heavy in cpu. It's lvmkd.
Output of dmesg:
LVM: Recovered Path (device 0x1f00a000) to PV 1 in VG 1.
LVM: Restored PV 1 to VG 1.
SCSI: Auto-Sense Unexpected Disconnect -- lbolt: 10216198, dev: 1f008000, io_id: 3a928d
SCSI: Auto-Sense Unexpected Disconnect -- lbolt: 10217734, dev: 1f008000, io_id: 3a9336
SCSI: Request Timeout; Abort Tag -- lbolt: 10221416, dev: 1f008000, io_id: 3a9352
SCSI: Auto-Sense Unexpected Disconnect -- lbolt: 10221916, dev: 1f008000, io_id: 3a97dd
SCSI: Auto-Sense Unexpected Disconnect -- lbolt: 10222919, dev: 1f008000, io_id: 3a98b2
LVM: vg[1]: pvnum=1 (dev_t=0x1f00a000) is POWERFAILED
LVM: Recovered Path (device 0x1f00a000) to PV 1 in VG 1.
LVM: Restored PV 1 to VG 1.
SCSI: Auto-Sense Unexpected Disconnect -- lbolt: 10253640, dev: 1f008000, io_id: 3aaf3c
SCSI: Auto-Sense Unexpected Disconnect -- lbolt: 10265763, dev: 1f008000, io_id: 3ab4fb

Line 5102 in /ux/core/kern/common/io/pat_psm.c: pat_get_olar_info - rc -1 slot_id -16187515

Thanks a lot.

[Steven E. Protter]

Shalom Carlos,

Based on the message, you have a disk that has either failed or is very close to failure.

I'd plan on a replacement, preceded by a backup if the data is still accessible. It seems to involve only one disk.

SEP

[Yogeeraj_1]

hi,

can you post the output of the following command?

ioscan -fknC disk


maybe one disk has already crashed!

revert

kind regards
yogeeraj

[Fabio Ettore · Accepted solution]

Hi,

yes, SEP is right and you should soon work on that disk. Anyway they seem two disks involved:

-----------
LVM: Recovered Path (device 0x1f00a000) to PV 1 in VG 1.

0x1f00a000 = c0t10d0
-----------

and

-----------
SCSI: Auto-Sense Unexpected Disconnect -- lbolt: 10216198, dev: 1f008000, io_id: 3a928d

1f008000 = c0t8d0
-----------

Also I already faced more times that message

Line 5102 in /ux/core/kern/common/io/pat_psm.c: pat_get_olar_info - rc -1 slot_id -16187515


and it is a sympthom that you should update or GSP or PDC firmware revision.

Good luck...

HTH.

Best regards,
Fabio

[Peter Godron]

Carlos,
the faulty disk in question should be c0t8d0.
1f008000

1f=major number for sdisk driver.
00=instance number for this controller
8=target

[Cheryl Griffin]

dev 1f008000 says it is c0t8d0.

Run some tests on the disk, such as diskinfo and dd. If you have diagnostics installed there are tests to exercise the disk.

pat_get_olar_info is an unrelated message that indicates you have a Lclass or Nclass, etc. that is running older firmware for GSP and PDC. You may want to look at updating those.

[Arunvijai_4]

Hi Carlos,

Try to run this command,

# dd if=/dev/rdsk/c0t8d0 of=/dev/null bs=64K count=1000

and see if this is successful.

-Arun

[Carlos_172]

ioscan -fknC disk
Class I H/W Path Driver S/W State H/W Type Description
======================================================================
disk 0 0/0/1/0.1.0 sdisk CLAIMED DEVICE HP DVD-ROM 304
/dev/dsk/c0t1d0 /dev/rdsk/c0t1d0
disk 1 0/0/1/0.8.0 sdisk CLAIMED DEVICE SEAGATE ST3146807LC
/dev/dsk/c0t8d0 /dev/rdsk/c0t8d0
disk 2 0/0/1/0.10.0 sdisk CLAIMED DEVICE SEAGATE ST3146807LC
/dev/dsk/c0t10d0 /dev/rdsk/c0t10d0
disk 3 0/0/1/0.12.0 sdisk CLAIMED DEVICE SEAGATE ST3146707LC
/dev/dsk/c0t12d0 /dev/rdsk/c0t12d0
disk 4 0/0/1/0.14.0 sdisk CLAIMED DEVICE SEAGATE ST3146707LC
/dev/dsk/c0t14d0 /dev/rdsk/c0t14d0
disk 5 0/0/1/1.2.0 sdisk CLAIMED DEVICE SEAGATE ST39103LC
/dev/dsk/c1t2d0 /dev/rdsk/c1t2d0
disk 6 0/0/2/0.0.0 sdisk CLAIMED DEVICE SEAGATE ST39103LC
/dev/dsk/c2t0d0 /dev/rdsk/c2t0d0
disk 7 0/0/2/0.2.0 sdisk CLAIMED DEVICE SEAGATE ST39204LC
/dev/dsk/c2t2d0 /dev/rdsk/c2t2d0

[Carlos_172]

seem another disk fails? (c0t10d0)?

SCSI: Resetting SCSI -- lbolt: 10470121, bus: 0
SCSI: Reset detected -- lbolt: 10470121, bus: 0
SCSI: Request Timeout; Abort Tag -- lbolt: 10473216, dev: 1f00a000, io_id: 3b6867
LVM: Path (device 0x1f00a000) to PV 1 in VG 1 Failed!
LVM: vg[1]: pvnum=1 (dev_t=0x1f00a000) is POWERFAILED
LVM: Recovered Path (device 0x1f00a000) to PV 1 in VG 1.
LVM: Restored PV 1 to VG 1.
SCSI: Auto-Sense Unexpected Disconnect -- lbolt: 10486342, dev: 1f008000, io_id: 3b7ef6
LVM: vg[1]: pvnum=1 (dev_t=0x1f00a000) is POWERFAILED
LVM: Recovered Path (device 0x1f00a000) to PV 1 in VG 1.
LVM: Restored PV 1 to VG 1.
LVM: Path (device 0x1f00a000) to PV 1 in VG 1 Failed!
LVM: vg[1]: pvnum=1 (dev_t=0x1f00a000) is POWERFAILED
SCSI: Request Timeout; Abort Tag -- lbolt: 10512416, dev: 1f00a000, io_id: 3b96fd

SCSI: Async write error -- dev: b 31 0x00a000, errno: 126, resid: 4096,
blkno: 126525444, sectno: 253050888, offset: 713035776, bcount: 4096.

SCSI: Async write error -- dev: b 31 0x00a000, errno: 126, resid: 4096,
blkno: 43983076, sectno: 87966152, offset: 2088996864, bcount: 4096.

SCSI: Async write error -- dev: b 31 0x00a000, errno: 126, resid: 8192,
blkno: 27734200, sectno: 55468400, offset: 2630017024, bcount: 8192.

SCSI: Async write error -- dev: b 31 0x00a000, errno: 126, resid: 4096,
blkno: 896652, sectno: 1793304, offset: 918171648, bcount: 4096.
LVM: Recovered Path (device 0x1f00a000) to PV 1 in VG 1.

[Fabio Ettore]

Hi Carlos,

yes, like I wrote in my post seem two disks involved:

*******************

-----------
LVM: Recovered Path (device 0x1f00a000) to PV 1 in VG 1.

0x1f00a000 = c0t10d0
-----------

and

-----------
SCSI: Auto-Sense Unexpected Disconnect -- lbolt: 10216198, dev: 1f008000, io_id: 3a928d

1f008000 = c0t8d0
-----------

*******************


Are they mirrored between them? Are they belong to the same VG? The good suggestion is to launch a dd to check and consider to open as soon as possible an HW call by HP. What about PE on these disks? You can see that by lvdisplay on logical volumes belong to them.

Good luck...

Best regards,
Fabio

[Carlos_172]

dd in c0t10d0 runs ok. dd in c0t8d0 hangs whith no response. Also, I run xstm over all disk in the system, and only c0t8d0 throws error. I call support. Thanks a lot gurus.

[Arunvijai_4]

Hello,

Yes, if "dd" for c0t8d0 hangs whith no response, you have defect.

-Arun

[Steven E. Protter]

I also like to confirm with cstm/mstm/xstm (take your choice).

If the disk is truly gone these tools won't see it. Your ouput indicates they will fail an stm disk excerize.

SEP

[Carlos_172]

Thanks a lot, folks!