Operating System - Tru64 Unix
1748039 Members
5175 Online
108757 Solutions
New Discussion юеВ

Re: mysterious crash alpha 4100

 
Michael Schulte zur Sur
Honored Contributor

mysterious crash alpha 4100

Hi,

we have repeated crashes on our alpha, running 5.1B, PK3 and so far HP was not able to tell us the reason though they exchanged a lot of parts, as my collegue told me.

I managed to create a crash dump and this is the panic string:
_panic_string: 0xfffffc00007e161c = "hardware restart"

I don't however know, whether this is from the panic or crash command. The strange thing, that we don't find any trace in errlog/messages.
Attached are two commands, the system suggested to run.

I would glad for any hint.

Michael
10 REPLIES 10
Ralf Puchner
Honored Contributor

Re: mysterious crash alpha 4100

the first time HP can not get the root cause of the problem. Are you sure you called the right service?

The attached console output is not valid for a software service to analyze the problem because this is the output of the CPU's normally analyzed by a hardware engineer.

"hardware restart" means someone pushed the "halt" button and restarted machine.

This call need deeper analyzis to check if it is hardware or software, so reopen the call within the HP support center!
Help() { FirstReadManual(urgently); Go_to_it;; }
Joris Denayer
Respected Contributor

Re: mysterious crash alpha 4100

Michael,

As Ralf said, this output must be investigated by an HW specialist.

In the beginning of your text, you see
"machine check while in PAL mode"
This means that the CPU is executing instructions from the PAL code. On that moment, it is even not working in Operating System "context".

Good Luck
To err is human, but to really faul things up requires a computer
Michael Schulte zur Sur
Honored Contributor

Re: mysterious crash alpha 4100

Ralf,

as I found out, the case was already under investigation by a hw analyst. I put the disks under stress and after less then two hours the machine haltet again.

thanks,

Michael
Ralf Puchner
Honored Contributor

Re: mysterious crash alpha 4100

Michael,

what was the result of the investigation of the HP support center? There is always a solution or case will not be closed (except customer agreed due to mystic whispering or fear of a ghost... ;-) )

So if case is under investigation or engineer must escalate it please standby or ask about status of call....

Help() { FirstReadManual(urgently); Go_to_it;; }
Michael Schulte zur Sur
Honored Contributor

Re: mysterious crash alpha 4100

Ralf,

I thought, I had posted info yesterday.
case# 1204166979
So far we managed to crash the machine with dd. But right now it is running stable. Will have to wait.

greetings,

Michael
Ralf Puchner
Honored Contributor

Re: mysterious crash alpha 4100

Michael,

why not waiting what engineer will find out? It is not quite fair to post parts of the messages here in the forum not waiting for end of diagnosis...
Help() { FirstReadManual(urgently); Go_to_it;; }
Michael Schulte zur Sur
Honored Contributor

Re: mysterious crash alpha 4100

Ralf,

fair? This call has been placed on May 12th!
I just wanted to see, if by any chance a forummer would get an idea before HP! ;-)

Michael
Ralf Puchner
Honored Contributor

Re: mysterious crash alpha 4100

Michael,

is it HP fault that you didn't ask for the final status of the call? Maybe the collegue is on holiday or ill not able to work closer on the problem. But we solved nearly all problems until yet... ;-)


Help() { FirstReadManual(urgently); Go_to_it;; }
Johan Brusche
Honored Contributor

Re: mysterious crash alpha 4100

M,

For repetitive "Machine Check" cases, where there is nothing in binary.errlog, I would advise to set the console in serial mode and capture the machine-check logout frames via a serial line of a PC and a terminal emulator. Besides of the info3/8 commands, 'show error' and 'sho fru' in console mode give also usefull information.

J.


_JB_