Summary: How to analyze Crash, to know if it is Hardware or Softw are ?

[Ashraf Baker <ABaker@xxxxxxxxxxx>]

Hi Admins
 I'm Sorry for being lat in summarizing  this article; I left the case under
monitoring, now it is closed. 

Thanks very much to all how respond... specially 
 Olle Eriksson, Emmanuel Bove, Lawrie Smith, Raj.,Stan Horwitz, Gavin
Kreuiter, alan, Mackun Roberto & Shaukat Riaz


The Original message was 

>  I have Digital 4100 with 5.1a operating system with patch kit 1.
> the system always crashes and gives different messages one of them is cpu0
> panic : uncorrectable machine error
> I want to ask how could I analyze crash files to know exactly what cause
> this problem,
> how could I know if the problem HW or SW is there diagnostics could I run
to
> check Server HW ?
>


 Actually cpu0 panic : uncorrectable machine error
was Hardware Error, The problem Found on memory after replacing memory
boards and monitoring for 2 weeks the server worked fine

 I thought there might be diagnostics CD to make Full Hardware Test, but you
can run test command from >>> to test HW.
 >>>test
  

 related to log files; we can make use of the following

/var/adm/messages                                                " for ASCII
error LOG"
/usr/sbin/uerf -t s:01-apr-2002 e:11-apr-2002     " for Binary error LOG"
/var/adm/crash/crash-data.nn                                 " Crash File"

not only uerf used to analyze /var/adm/binary.errlog log file also we can
use uerf/dia/ca 
we can using Compaq Analyzer (ca),"ca x analyse". 
or dia -R 

 kernel debugging can be used to collect more information refer to kernel
debugging book
2 files are copied "other than crash-data.nn "used in analyzing crash in
/var/adm/crash/
kernel and crash dump; "But I think it is complex way"


 Thanks again to all how responded 


Best regards

ENG. Ashraf Baker
Technical Support Engineer
DANA Information Systems
 
Tech. Dept. : +962  6  5169080
FAX            : +962  6  5160663
Mobile         : +962 79 5692240