| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 2022.1 | more info | BACHUS::DEVOS | Manu Devos NSIS Brussels 856-7539 | Fri Apr 25 1997 09:04 | 10 |
| Hi Paul,
Please, give us more information on your hardware configuration.
Did you change the SCSI ID of the controllers ? Is the local disk
really local or did you try to place it on the shared bus ????
A drawing of the config is welcome ...
Manu.
|
| 2022.2 | check basic-dma-window-size | FOUNDR::STRICKER | Enterprise Systems Engineering, Salem NH | Fri Apr 25 1997 10:56 | 20 |
| Paul,
> show device and test commands at the console level produce errors
> associated with kzpsa2 and kzpsa3 which are connected to HSZ50 disks
> in SW800 cab. This is the case from both system consoles.
Did you check the HSZ50s status? Are the batteries in a GOOD state,
meaning are they fully charged?
> load of /etc/init failed, errno 2
> load program?
> **** what is it asking for?
Did anyone change the sysconfigtab file? This looks like the same problem
that I was seeing on 4100s with >1GB memory when I tried to bootup UNIX
with basic-dma-window-size=512. The workaround was to change this value to
either 0 (zero) or 3072 (number of MB of memory).
-Jerry
|
| 2022.3 | Also slot restrictions | NNTPD::"cherkus@buff.zk3.dec.com" | Dave Cherkus | Fri Apr 25 1997 17:01 | 9 |
| Also search this notesfile using the keyword '4100' for a discussion
on the proper slots to use for memory channel in 4100 for DU 3.2x.
One piece of advice is: don't expect your system to perform reliably
if the console is complaining about your storage devices. This
should not be ignored.
Dave
[Posted by WWW Notes gateway]
|
| 2022.4 | Could be this... | NNTPD::"mcdonald@decatl.alf.dec.com" | John McDonald | Mon Apr 28 1997 19:15 | 86 |
| Paul,
This may or may not be your problem, but it's worth looking into:
Author : MARCI R POTTER
User type : DBA
Location : USTIMA
Vaxmail address : CSC32::POTTER
Copyright (c) Digital Equipment Corporation 1997. All rights reserved.
+---------------------------+TM
| | | | | | | |
| d | i | g | i | t | a | l | TIME DEPENDENT BLITZ
| | | | | | | |
+---------------------------+
BLITZ TITLE: Alpha Server 4100 - SCSI Bus and Related Errors
DATE: April 24, 1997
AUTHOR:Ted Gent TD #: 2274
DTN:223-6530
ENET:POBOXA::GENT CROSS REFERENCE #'s:
DEPARTMENT:SBU Engineering (PRISM/TIME/CLD#'s)
INTENDED AUDIENCE: U.S/EUROPE/GIA PRIORITY LEVEL: 2
(1=TIME CRITICAL,
2=NON-TIME CRITICAL)
=====================================================================
Subject : Alpha Server 4100 - SCSI Bus and Related Errors
AlphaServer 4100 - Errors on SCSI Busses
1. Problem:
Test coverage on the B3040 'Horse' module has been found marginal in
identifying problems in the 'Scatter-Gather' mapping. The result is that
some modules have been shipped which may have latent defects.
2. Susceptibility:
Scatter-Gather mode use varys depending on the adapter type. The NCR810
based controllers - KZPAA and the On-Board SCSI controller for the CD-ROM
always use Scatter-Gather mode. Unix uses Scatter-Gather mode for other
SCSI controllers if the memory configuration exceeds 1 GByte.
3. Symptoms:
Symptoms of the problem will vary, some of the problems seen include:
a) Excessive SCSI errors on NCR810 based devices
b) SCSI (CAM) errors reported during the system boot
c) System crashes (panics) with Bus Faults
d) Performance degradation accessing disk subsystems or CDROM
e) Occasional Power Up Self Test failures indicating an IOD failure
f) Failure to initialise graphics correctly
g) Moving all of the PCI adapters to one (of the) PCI bus makes the system
function properly.
4. Fault Finding:
If you are involved in trouble shooting a system where you are experiencing
any of the symptoms listed above, the B3040 may be the problem module.
5. Prevention:
A screen is being implemented in Stage 1 and Stage 2 Manufacturing to
remove defective B3040 modules. The screen went into effect in April. This
means that any systems currently in the field may have this problem
6. Severity:
Screening indicates that this problem may be seen on as high as 10% of the
population of B3040's
tg 4/22/97
John McDonald
Atlanta CSC
[Posted by WWW Notes gateway]
|
| 2022.5 | fixed- Thanx Jerry! | SANITY::PCUMMINGS | The perfect democracy | Tue Apr 29 1997 13:10 | 16 |
| the no-boot problem was fixed by booting the DU o.s CD in sys mgmt mode
bringing us to the Unix Shell prompt, mounting the SWXCR HW raid root
disk as /mnt and using 'ed' to edit the /etc/sysconfigtab file. the
problem was that reinstalling the TCR*100 subsets modifies the
sysconfigtab file with a bogus basic-dma-window-size=512. Once you
reboot with this value, you're hozed. Using 'ed' we reset the value to
3072 the init and Memory Channel Hub errors went away and system booted
fine.
The CAM & HSZ5 errors during reboot remained until we powered down the
SW800 disk cab (which has HSZs too). Strange thing is the other
cluster member, didn't get these CAM & HSZ errors.
thanx
/paul
|