Re: Kernel load reporting

[Matt Pavlovich <mpav algx net>]

I have verified that the number of work units performed does not even
come close to what should have been done-- looks like VM/Scheduler
issue.  Bugzilla ID: 101465

mpstat shows some interesting numbers:
[root daldev35 root]# mpstat -P ALL
Linux 2.4.21-1.1931.2.349.2.2.entbigmem (daldev35)  08/01/2003
10:40:01 AM  CPU   %user   %nice %system   %idle    intr/s
10:40:01 AM  all    0.01   96.19    3.23    0.57    112.52
10:40:01 AM    0    0.02   96.57    2.63    0.55    112.52
10:40:01 AM    1    0.00   97.24    2.18    0.61      0.00
10:40:01 AM    2    0.01   96.72    2.75    0.55      0.00
10:40:01 AM    3    0.00   97.32    2.09    0.62      0.00
10:40:01 AM    4    0.01   95.92    3.52    0.58      0.00
10:40:01 AM    5    0.00   96.95    2.55    0.53      0.00
10:40:01 AM    6    0.02   94.28    5.15    0.58      0.00
10:40:01 AM    7    0.03   94.51    4.94    0.55      0.00

Suck!

Matt Pavlovich

> Message: 1
> Date: Thu, 31 Jul 2003 16:35:56 -0500
> From: Matt Pavlovich <mpav algx net>
> Subject: Re: Kernel load reporting
> To: taroon-beta-list redhat com
> Organization: Allegiance Telecom, Inc.
> Reply-To: taroon-beta-list redhat com
> 
> One odd tidbit-- I have not been able to actually verify that seti is
> processing 100%.  The result stats from yesterday don't reflect running
> 8 seto instances on 2Ghz XEONs.  I am working on verifying that on my
> end.
> 
> If it is really only processing at a few percentage points, then there
> is a real serious issue. 
> 
> Matt Pavlovich