SCSI: Write error

[SM_3]

hp ux 11.00
b180 workstation


SCSI: Write error -- dev: b 31 0x006000, errno: 126, resid: 1024,
blkno: 1362792, sectno: 2725584, offset: 1395499008, bcount: 1024.
SCSI: Resetting SCSI -- lbolt: 61363, bus: 0
SCSI: Reset detected -- lbolt: 61363, bus: 0
SCSI: Unexpected Disconnect -- lbolt: 65511, dev: cb003000, io_id: 1494


SCSI: Read error -- dev: b 31 0x006000, errno: 126, resid: 2048,
blkno: 8, sectno: 16, offset: 8192, bcount: 2048.

SCSI: Read error -- dev: b 31 0x006000, errno: 126, resid: 2048,
blkno: 8, sectno: 16, offset: 8192, bcount: 2048.
LVM: Recovered Path (device 0x1f006000) to PV 0 in VG 0.
LVM: Restored PV 0 to VG 0.

is the root disk (c0t6d0) on it's way out?

[John Dvorchak]

If that is the root disk then I'd say yeah.... Hope you have a mirror or at the very least an recent Ignite tape.

Good luck

[Jeff Schussele]

Well, it's certainly an indication that it (c0t6d0) is having problems with several blocks.
I'd run the exerciser from the STM suite to test it.
Or even a dd to /dev/null as a test.

If all that's happened is a bad spot has popped up, you may be able to mirror it with another drive, reduce it from the mirror & reformat it with mediainit.

Rgds,
Jeff

[SM_3]

this disk is not a 24x7

it's up and running 9am-5pm and that's it

how come it's dying?

[Patrick Wallek]

I would agree. The disk does appear to be on its last legs.

Since you are getting both read and write errors, I'd say it's time to start planning how and when to replace that disk.

[S.K. Chan]

Yes, it's possible ! You can run cstm "exercise" tool to confirm.
# cstm
cstm> map
cstm> sel dev
cstm> exercise
Run "map" to check for completion. Then ..
cstm> eal
cstm> efl

[Michael Tully]

Being a hardware component, there is no actual way to tell the lifetime of a component. Hardware failures can be unpredictable, you cannot always tell of a failure, but these types of messages are an indicator that the hardware may fail very soon. Plan now for the component replacement.

[S.K. Chan]

I hope this is not a machine that you shutdown after 5pm (just like turning off a PC) when you say it's not 24x7 and it's certainly not uncommon for hard disk to fail.

[John Dvorchak]

To answer your question "How come it's dying" because it was powered on.

As the old saying goes, there are two kinds of System Admins, those that have lost a disk and those that are going to.

Disks are electromechanical devices and subject to wear and tear. Like any mechanical device it will wear out. And like any electronic device, it will fail sooner or later.

[SM_3]

except 24x7 servers

this is the first dying disk I've had

[Lasse Gandlose]

Do not worry about that ! you will get another one sooner or later! :-) this is common but you are right that it is mostly seen when you reboot or install new firmware.

[SM_3]

I have another problem now

the root disk was mirrored

now the mirror is complaing about the same problem

grrrreat!

nevertheless I can still boot up in run level 3

thanks for the help

[Steven E. Protter]

lbolt means the disk inidicated is going to go. You need to get an ignite backup (make_tape_recovery) as soon as possible, replace the disk and get on with life. Don't ignore it, it won't go away.

HP hardware always replaces hardware generating these errors.

Possible causes:

Drive cage(if applicable).

SCSI I/O subsystem, on some systems its connected to the GSP card.

scsi cable

scsi card.

scsi device.

SEP

[Eugeny Brychkov]

SM,
look at the back of the B180 system. There're at least one SCSI connector (SE). Please make sure there's (or there're in case B180 has more than one SCSI adapters) correct terminator attached: SE for SE output, UWSE for UWSE output.
If they're ok, then it's good idea to make ignite image of root disk and replace it (at least try replacing to see if problem will go away)
Eugeny