K370 SCSI Reset Errors after OS Upgrade to 11i

[Dan Matlock_1]

Question: Is there dependancies that I have missed after a 'cold install' of HP-UX 11i. I upgrade this K370 from 11.00 to 11i. At first glance I thought I have a physical disk failure pending, but I just not sure now. Nothing else physically has changed except the OS and Patches.
GOLDAPPS11i B.11.11.0509.429 Applications Patches for HP-UX 11i v1, September 2005
GOLDBASE11i B.11.11.0509.429 Base Patches for HP-UX 11i v1, September 2005
HWEnable11i B.11.11.0509.430 Hardware Enablement Patches for HP-UX 11i v1, September 2005
HPUXBase64 B.11.11 HP-UX 64-bit Base OS

EMS implies that there might be a 'firmware' issue.

Kernel was 64bit prior.

Since OS upgrade and patches, I now get sloooow performance do to the following in the syslog.log:

May 14 16:33:50 nxphrd01 vmunix: SCSI: Resetting SCSI -- lbolt: 8386272, bus: 0
May 14 16:33:50 nxphrd01 vmunix: SCSI: Reset detected -- lbolt: 8386272, bus: 0
May 14 16:33:50 nxphrd01 vmunix: SCSI: Request Timeout; Abort Tag -- lbolt: 8388672, dev: cb006000, io_id: 73d05
May 14 16:34:14 nxphrd01 vmunix: SCSI: First party detected bus hang -- lbolt: 8389972, bus: 0
May 14 16:34:14 nxphrd01 vmunix: lbp->state: 5020
May 14 16:34:14 nxphrd01 vmunix: lbp->offset: f0
May 14 16:34:14 nxphrd01 vmunix: lbp->uPhysScript: 780000
May 14 16:34:14 nxphrd01 vmunix: From most recent interrupt:
May 14 16:34:14 nxphrd01 vmunix: ISTAT: 01, SIST0: 00, SIST1: 00, DSTAT: 84, DSPS: 00000010
May 14 16:34:14 nxphrd01 vmunix: lsp: 0000000000000000
May 14 16:34:14 nxphrd01 vmunix: lbp->owner: 0000000040f47500
May 14 16:34:14 nxphrd01 vmunix: bp->b_dev: 1f006000
May 14 16:34:14 nxphrd01 vmunix: scb->io_id: 73bc7
May 14 16:34:14 nxphrd01 vmunix: scb->cdb: 28 00 00 07 da 00 00 00 40 00
May 14 16:34:14 nxphrd01 vmunix: lbolt_at_timeout: 8389372, lbolt_at_start: 8389372
May 14 16:34:14 nxphrd01 vmunix: lsp->state: 10d
May 14 16:34:14 nxphrd01 vmunix: scratch_lsp: 0000000040f47500
May 14 16:34:14 nxphrd01 vmunix: Pre-DSP script dump [0000000044013030]:
May 14 16:34:14 nxphrd01 vmunix: 78347000 0000000a 78350800 00000000
May 14 16:34:14 nxphrd01 vmunix: 0e000004 00780540 80000000 00000000
May 14 16:34:14 nxphrd01 vmunix: Script dump [0000000044013050]:
May 14 16:34:14 nxphrd01 vmunix: 870b0000 007802d8 98080000 00000005
May 14 16:34:14 nxphrd01 vmunix: 721a0000 00000000 98080000 00000001
May 14 16:34:14 nxphrd01 vmunix: SCSI: Resetting SCSI -- lbolt: 8390072, bus: 0
May 14 16:34:14 nxphrd01 vmunix: SCSI: Reset detected -- lbolt: 8390072, bus: 0

[A. Clay Stephenson]

dev: cb006000

The first thing that I would do is an lsdev and determine which character device corresponds to major device number 203 (cb hex). This could well be a tape drive (c0t6d0). If so, I would disconnect this device and see if your problems disappear.

[Bill Hassell]

Since this is a cold install, why use 2005 patch bundles. The latest bundles are Dec 2006:

http://www1.itrc.hp.com/service/patch/releasePage.do?BC=main|releaseIndexPage|&releaseId=5013-6487

But before you install HP-UX, bring the K-box up to the latest firmware level. Unfortunately, the firmware has been moved to a different site and they apparently lost all the firmware (none available for the dozen models I tried). Did anyone manage to save K-class firmware somewhere?

[Dan Matlock_1]

Thanks for the input. I choose 2005 only because the production peoplesoft server we have is at that patch level, and it was requested that this DEV server be brought upto the same OS/Patch level to match it. Unfortunetly for me its a OLD K-box. I did surf the web for firmware updates and have yet to find any. Find it hard to believe that a HW issue spawned only after a OS software update, so was leaning towards firmware/software patch issue..

[A. Clay Stephenson]

I'll be surprised if firmware that worked for 11.0 will not also work for 11.11 with the same hardware but, in any event, the Force is with you. I was wise enough to save all the versions of firmware for my old boxes before they disappeared. I have PF_CBHK4201 (42.01 is the latest for Kx70's and Kx80's). It's too large to attach in one chunk so I split it for you into 3 chunks that you can cat back together.

Here is pfaa.

[A. Clay Stephenson]

Here is the 2nd part, pfab.

[A. Clay Stephenson]

and here is the 3rd part, pfac.

Download all 3 files (using the names I suggested) to the /tmp directory on and then "cat pf* > PF_CBHK4201". You then unshar the file "sh PF_CBHK4201" to creat a firmware file, CBHK4201.frm, and the documentaion, CBHK4201.text. Read the .text file for instructions on making the bootable tape image.

[Torsten. · Accepted solution]

It's still here:

ftp://ftp.itrc.hp.com/firmware_patches/hp/cpu/PF_CBHK4201

and

ftp://ftp.itrc.hp.com/firmware_patches/hp/cpu/PF_CBHK4201.txt

[Marcel Burggraeve]

There's also firmware for quite some models at the following site :
ftp://ftp.itrc.hp.com/archived_patches/firmware_patches/hp/cpu/

[Marcel Burggraeve]

Forgot my reply regarding the SCSI errors.
This might sound weird but are you totally sure everything is terminated like it should ?
I've seen loads of systems in the field without proper termination which seemed to work without any problem for a long ( or short ) period of time and started to act weird all of a sudden.
Weird being crashes, unable to boot, SCSI errors, slow boots etc etc.