On Thursday, October 30, 2003, at 12:48  AM, Jochen Kunz wrote:

>> The Service Handbook says this is CHASSIS_FATAL_ERR_WRITING_EEPROM:
>> "Fatal Error occurred writing to EEPROM."
> Did this happen when you tried booting the kernel?
> The kernel and the boot loader should not write to the NVRAM or
> STABLE_STORAGE. It would be bad if they do.

No, this was during hardware self-test, before any boot attempts.

>> Is this something that can be fixed by replacing the EEPROM?
>> If so, does anyone know what kind of EEPROM chip it is?
> Maybe somthing got loose during transportation? Somtimes this can be
> cured by opening the machine, reseating all cards and pressing on all
> socketed chips.

I found the EEPROM on the motherboard and made sure it was firmly in 
the socket. No help. Then I took it out, peeled off the label (which 
has some kind of hardware identification number on it), and was happy 
to find a part number. Tracking it down, I was able to determine the 
specific kind EEPROM it is and order a replacement. We'll see if that 
helps.

> Note that the J282 is a PA8000 i.e. 64 bit machine. (Does it have 
> single
> or dual CPU? AFAIK J28_2_ indicates dual.)  Maybe it will crash very
> early during boot, maybe it will just run in 32 bit mode. I have no
> PA8xxx machine to test, but I will get a C240 in some weeks.

It is indeed a dual-processor system. If the new EEPROM solves the 
self-test problem, I'll be able to test on this machine.

Jochen, I also have a 712/60 I can test on, if that will be of any help.

Kirill