asked on

solaris error.

Hallow,
/var/adm/messages show the following. uname -r shows 5.7
"unix: [AFT2] errID 0x002e7772.72894422 PA=0x00000000.0c927230
E$tag 0x00000000.08400192 E$State: Shared E$parity 0x04
unix: [AFT2] E$Data (0x00): 0x0014a174.0014a17c
[AFT2] E$Data (0x08): 0x0014a184.0014a18c
unix: [AFT2] E$Data (0x10): 0x0014a190.0014a198
unix: [AFT2] E$Data (0x18): 0x0014a19c.00138f10
unix: [AFT2] E$Data (0x20): 0x00112b00.00112c00
unix: [AFT2] E$Data (0x28): 0x00138fd4.00112c50
unix: [AFT2] E$Data (0x30): 0x00112c5f.00112c6e *Bad* PSYND=0xff00
unix: [AFT2] E$Data (0x38): 0x00112c80.00138fd8
unix: [AFT2] errID 0x002e7772.72894422 PA=0x00000000.0c927230
E$tag 0x00000000.08400192 E$State: Shared E$parity 0x04
unix: [AFT2] E$Data (0x00): 0x0014a174.0014a17c
unix: [AFT2] E$Data (0x08): 0x0014a184.0014a18c
unix: [AFT2] E$Data (0x10): 0x0014a190.0014a198
unix: [AFT2] E$Data (0x18): 0x0014a19c.00138f10
unix: [AFT2] E$Data (0x20): 0x00112b00.00112c00
unix: [AFT2] E$Data (0x28): 0x00138fd4.00112c50
unix: [AFT2] E$Data (0x30): 0x00112c5f.00112c6e *Bad* PSYND=0x0400
unix: [AFT2] E$Data (0x38): 0x00112c80.00138fd8
unix: [AFT3] errID 0x002e7772.72894422 Above Error is in User Mode
and is fatal: will reboot......
plz.explain what the error stands for and how to debug it.
Thanx,
BY

ASKER CERTIFIED SOLUTION

liddler

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

byerabati

ASKER

Thanx liddler,
how can i know the error is due to bad memory or cpu.how to read the error.

Thanx

liddler

I don't know, any error like this I pass to Sun, as they support the hardware on all of my systems.

You could have a look at the output of
/usr/platform/`uname -i`/sbin/prtdiag -v
to see if that reports any hardware problems

Also have a look at:
http://forums.sunhelp.org/archive/index.php/t-372
http://sunportal.sunmanagers.org/pipermail/summaries/2004-January/004790.html
http://sunportal.sunmanagers.org/pipermail/summaries/2002-January/000662.html

Matt_Avery

liddler is correct. It's an E-Cache ("E$") error, which means that there is some kind of hardware problem with the CPU cache memory.

If you send that data to Sun, they will most likely accept it as proof of a hardware failure and fix it for you if you have a support contract or are covered under warranty.

SOLUTION

SumeshDaftary

membership

This solution is only available to members.

To access this solution, you must be a member of Experts Exchange.

Start Free Trial

colinbo

Just one point to clarify regarding Sun's support policy on the E-cache errors having spent the last 6 years in a Sun Solution Centre (before I was laid off and my job got out-sourced to low-cost engineers in India):
If this error has just been a one-off event then Sun Support is likely to say "call us next time it happens again on that CPU *then* we'll replace it."

If this has happened more than once already then Sun should replace the CPU module.

Believe it or not, transient e-cache errors can be caused by cosmic rays, however the chances of naturally occurring causes of the error hitting the same CPU twice are almost zero. Hence the "replace after 2 hits" policy.

The newer UltraSPARC III CPUs now in all new Sun h/w have ECC functions to allow single bit error correction on transient e-cache errors and the latest policy on replacing those CPUs is >24 errors in 24 hours, or immediately on any double bit errors.

A very detailed scientific document on the subject of naturally occurring causes of memory faults can be found at the following URL:
http://www.research.ibm.com/journal/rd/421/ziegler.html

Don't ask me how I came to have this URL, I bookmarked it years ago and it's thankfully still there, otherwise people wouldn't believe me when I mention cosmic rays. HTH

- CB

colinbo

Hey liddler: what about some points for the info about cosmic rays? :-)