Alphaserver 1000A & ISP1020 SCSI chip

From: Christophe Wolfhugel <wolf_at_pasteur.fr>
Date: Wed, 26 Jun 1996 14:37:11 +0200 (MET DST)

Greeting.

We have a 1000A 4/266 server which is regularly crashing on a machine check
error. The french DEC support is not really efficient, except at changing
CPU and mother boards.

Digital UNIX is 4.0, SRM revision 4.5-72, PALcode : OSF 1.45, ARC 4.47
(update from the 3.6 firmware update at ftp.digital.com).

Hardware : internal ISP1020A connected to the CD, KZPSC with two ports
connected to the internal StorageWorks shelf ; ethernet and PCI FDDI.

First message :

Jun 26 14:13:44 schnaps vmunix: pci2000 at pci0 slot 8
Jun 26 14:13:44 schnaps vmunix: isp0 at pci2000 slot 0
Jun 26 14:13:44 schnaps vmunix: isp0: QLOGIC ISP1020A
Jun 26 14:13:44 schnaps vmunix: isp0: INFO: NVRAM parameters invalid, using driver defaults
Jun 26 14:13:44 schnaps vmunix: isp0: Firmware revision 2.10 (loaded by console)
Jun 26 14:13:44 schnaps vmunix: isp0: StorageWorks signal reporting enabled
Jun 26 14:13:44 schnaps vmunix: isp0: StorageWorks SHELF_OK signal transition reported
Jun 26 14:13:44 schnaps vmunix: scsi0 at isp0 slot 0

Should we care about the INFO message ? If yes is there a way to correct this
from the OS or should we try to get some support diskette ?

Also :

[...]
Jun 26 14:13:45 schnaps vmunix: kernel console: cirrus0
Jun 26 14:13:45 schnaps vmunix: dli: configured
Jun 26 14:13:46 schnaps vmunix: WARNING: too many Processor corrected errors det
ected on cpu 0. Reporting suspended.
Jun 26 14:13:46 schnaps vmunix: Machine Check error corrected by processor

>From the UERF :

EVENT CLASS ERROR EVENT
OS EVENT TYPE 100. CPU EXCEPTION
SEQUENCE NUMBER 1.
OPERATING SYSTEM DEC OSF/1
OCCURRED/LOGGED ON Wed Jun 26 14:12:40 1996
OCCURRED ON SYSTEM schnaps
SYSTEM ID x0006001B
SYSTYPE x00000000

----- UNIT INFORMATION -----

UNIT CLASS CPU

Nearly all the hardware has been changed, and I had several other Alpha 1000's
(not 1000A and running DU 3.2) with similar problems : regularly crashing
on machine check errors regardless of processor card and motherboards and
memory which have been changed several times).

I would be tempted to think that there is some sort of microcode or fimware
which is not up to date somewhere... but nor I nor DEC have an idea on where.

-- 
Christophe Wolfhugel  -+- SIS, Institut Pasteur, Paris
Boulot : wolf_at_pasteur.fr,  $HOME : wolf_at_schnok.fr.net
Received on Wed Jun 26 1996 - 15:06:55 NZST

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:46 NZDT