We have 4 DEC Alpha dual-processor 2100A's running DU 4.0b...three of
them have 'lost' one of the CPUs without any notice to the console or
the error log.
Has anyone seen this bahavior or have any idea what would cause it??
I apologize for the length of this message in advance!
BACKGROUND:
We have four rack-mounted Alpha 2100A 5/300 systems. They are
identically configured:
2x 21164 300 MHz CPUs
128 MB ram
1 RZ29B
KZPAA SCSI controller
1 DECchip 21040-AA ethernet controller
DWPVC-BA PCI-VME host bus adapter
1 TLZ09 4 mm DAT tape drive
Third-party cards include:
1 SCRAMNet+ PCI reflective memory card (one machine only)
1 DCC5-P (clock card) - all machines
1 Bit-3 PCI card - all machines
1 VMIC VMIPCI-5588 Reflective Memory card - all machines
The kernels on these systems are configured to communicate with a VME
chassis and have several third party device drivers installed.
PROBLEM:
While troubleshooting a piece of code that required multple CPUs, I
discovered that only one CPU was showing up on 2 of the 4 systems.
I shut one system down and from the console prompt I could see both
CPU cards, but one of them had question marks in the Module ID and
both cards had a "F" under Status.
P00>>> show config
Component Status Module ID
CPU 0 F B2040-BA DECchip(tm) 21164-5
CPU 3 F B2040-BA DECchip(tm) ?????-?
P00>>> show fru
Rev Events logged
Option Part# Hw Sw Serial# SDD TDD
CPU0 B2040-BA B1 34 KA708TYVWA 00 01
CPU3 B2040-BA B1 34 KA708TYVVJ 00 01
RESOLUTION:
I called hardware support and they recommended the following action.
It worked and since then we had the problem occur on a THIRD system!
P00>>> clear_error all
P00>>> init
P00>>> show fru (write down the info)
P00>>> set mode diag
P00>>> build cpu3 b2040-ba ka01234567 b1 34
P00>>> init
If the question marks are still there, cycle the power and boot.
This seems to fix the problem, but it it likely to happen again. I
would like to understand how to prevent it before these systems go
into production!
Thanks in advance for your help!
Brenda Ramsey
Hughes Training, Inc.
Arlington, Texas
Received on Tue Nov 11 1997 - 22:55:27 NZDT