hdisk1: Error log analysis indicates a hardware failure

[Jason Heirtzler]

One of these errors poped up on the console of my 320 this
morning. The error message (and the output of 'errpt -a')
isn't very enlightening, unless you know how to interpret
the SCSI sense data.

Maybe someone can decipher the error and tell me if this is
simply a bad block (lvm handles this for you, right?) that can
be mapped out, or it indicates a more ominous failure, that
requires me to replace the disk. Also I had one (and only one)
error like this on hdisk0, which would cast suspicion on the
SCSI controller, or cabling perhaps.

Calling 1-800-IBM-SERV is certainly an option, but I've been
delaying this for fear that the support people may not fully
understand the error and just say "let's replace the disk."

Here's the relevant output of 'errpt -a'

The release is AIX 3.1.7

----------------------------------------------------------------------

ERROR LABEL: DISK_ERR4
ERROR ID: B360B018

Date/Time: Apr 20 15:49:02
Sequence Number: 19706
Machine Id: 000149753100
Node Id: squid
Error Class: H
Error Type: TEMP
Resource Name: hdisk1
Resource Class: disk
Resource Type: 320mb
Location: 00-01-00-10
VPD:
Device Specific.(00000
Manufacturer................IBM
Machine Type and Model......0661-371
Serial Number...............00076960
Device Specific.(193X0928 098090158 00022293X2355 843839
93X2488 843841

Error Description
DISK FAILURE

Probable Causes
DASD MEDIA
DASD DEVICE

Failure Causes
DASD MEDIA
DISK DRIVE

Recommended Actions
PERFORM PROBLEM DETERMINATION PROCEDURES

Detail Data
SENSE DATA
0601 0000 0A01 14B8 2000 0000 0000 0000 0102 0000 7000 0100 0000 0018 0000 0000
0300 0080 0003 0000 01C5 0000 FFFF FFFF 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 0005 0006 8AC0

--------------------------------------------------------------------------

Jason Heirtzler (617) 353-2780 j...@bu-pub.bu.edu
Information Technology Boston University ..!bu.edu!bu-pub!jdh

"And now the sound of John Denver being strangled.."

[Randy Swanberg]

The sense data indicates a 'Peripheral Device Write Fault'. It sounds
worse than it
really is. It is a recovered error caused by either a temporary arm
electronics
error or a sector overrun.

_____________________________________________________________________________

Randy Swanberg VM VNET: SWAN@AUSVMQ
PSP/AIX Operating System Arch/Dev PSP LAN: sw...@swan2.austin.ibm.com
Office 8E-021/905, ZIP 9582 Ext: 8-3363 T/L: 678-3363
_____________________________________________________________________________

[Torben N. Rasmussen]

We had a similar problem on one of our RS/6k's. This message appeared
periodically over several months, until the disk died and had to be
replaced.