[Orca-users] [Fwd: Re: ks_system_pages.pp_kernel]

[Duncan Lawie]

From: Duncan Lawie <duncanl at demon.net>

Hi,

I tried this one on the gurus themselves - known bug.  I'm getting the
ridiculous output on my database server and also on another system. 
When I close down the application on the second system the figures all
make sense.

Cheers,
Duncan.

-------- Original Message --------
Subject: Re: ks_system_pages.pp_kernel
Date: Thu, 9 Sep 1999 10:14:53 -0700 (PDT)
From: Rich Pettit <richp at foglight.com>
Reply-To: Rich Pettit <richp at foglight.com>
To: se-feedback at chessie.eng.sun.com, duncanl at demon.net

> I'm getting values for pp_kernel of over 4G from some of my systems.  I
> have two Ultra-250s, one of which produces a sensible figure.  They each
> have the same value for total pages, which is reasonable as each has the
> same amount of real memory.
> 
> Am I doing something silly?
> 
> I'm using SE 3.1preFCS.  Putting the gigabit ethernet fixes in didn't
> make any difference.
> 
> included:  /etc/release, memory size, psrinfo, uname -a and script
> output for each box, along with my se script.

There is a known bug in the computation of pp_kernel on systems using
Oracle and shared memory. I think that's the deal. pp_kernel will be
reported absurdly high.

Rich
---- Richard Pettit                  richp at foglight.com
---- Chief Performance Architect
---- Foglight Software, Inc.       http://www.foglight.com

--------------------------- ONElist Sponsor ----------------------------

BETA TEST expertcity.com's new online, real-time, technical support service!
   Let our live experts help you with your computer now. IT'S FREE! ! !
         <a href=" http://clickme.onelist.com/ad/expertcity8 ">Click Here</a>

------------------------------------------------------------------------