1753253 Members
4213 Online
108792 Solutions
New Discussion юеВ

Re: K580 HPMCs

 
Walt Stupinski
Occasional Advisor

K580 HPMCs

PIMs indicate Processor Memory Bus Parity errors (7D05 in all processors). All PIMs also show 7F00 (Carrier 0, Pair 0 - Am I right so far?). Each carrier has (4) pairs of 256M.

Removed that memory pair (replacing it with the pair from the highest slot, and now get a daily HPMC w/Fault 7D06 Write Bomb Fault HPMC - Earlier write ti memory had a bus parity error).

Is the Write Bomb related, or could it be another trouble?

Running 4 processors.

Walt

5 REPLIES 5
S.K. Chan
Honored Contributor

Re: K580 HPMCs

7D06=HPMC memory address fault
The "write bomb" fault makes me want to say you probably got a CPU fault .. but what you should do is retrieve the HPMC information to find out exactly what the fault is.
To do this shutdown the machine. After that turn the key to "SERVICE" position.
Then enter CTRL-B at console and RS to reset it. Interrupt the bootup then goto the "service menu" (I think), and enter

pim

Take note and post the error here (or call HP response center).
Walt Stupinski
Occasional Advisor

Re: K580 HPMCs

Thanks for the response. My book shows 7D06 as a Write Bomb Fault.

Anyway that was the indication on the panel. PIM seems to show differently though:

CPU 0:
CBF0 500A 5408 5508 CBFB

CPU 1:
(no Codes Logged)

CPU 2:
CBF0 5208 5408 5508 CBFB

CPU 2:
CBF0 5308 5408 5508 CBFB

Walt
Robert_Jewell
Honored Contributor

Re: K580 HPMCs

The code 500A on CPU 0 is actually a rare code. Its reffered to as an unknown bus error. The other codes that end in 8 simply say to check other modules. The fact that CPU1 didnt report anything is interesting as well. I would swap 0 and 1 to see if that changes anything. You could also just remove CPU 0 (move over CPU2) and again see if that shows anything different.

-Bob
----------------
Was this helpful? Like this post by giving me a thumbs up below!
Patrick Wessel
Honored Contributor

Re: K580 HPMCs

Walt,
I'm afraid that replacing the memory wasn't the right step. Along with a runway bus parity error is a chassis code of 7Fxx worthless. But the write bomb fit's into that scenario.
You can't troubleshoot a runway bus parity errors just by chassis codes. Could you please provide the whole PIM data from the last two HPMCs?
There is no good troubleshooting with bad data
Geno Church
Frequent Advisor

Re: K580 HPMCs

I see you posted no codes for cpu1. If it didn't report then it would appear to me that he is causing the HPMC. I would recommend replacing cpu1.

Geno
Real Eyes Realize Real Lies