Load avg issue

[Andrei Pascal]

On Fri, 2007-04-13 at 02:09 -0400, nilesh vaghela wrote:
> when the server is normal the top say
> 
> 12:01:26  up 91 days, 20:07,  3 users,  load average: 1.71, 1.23, 1.66
> 172 processes: 166 sleeping, 4 running, 2 zombie, 0 stopped
> CPU states:  cpu    user    nice  system    irq  softirq  iowait    idle
>            total   57.0%    0.0%    5.6%   0.0%     0.0%    9.8%  127.2%
>            cpu00   24.1%    0.0%    3.3%   0.0%     0.0%    5.9%   66.4%
>            cpu01   32.8%    0.0%    2.3%   0.0%     0.0%    3.9%   60.7%
> Mem:  1024780k av, 1007700k used,   17080k free,       0k shrd,   38324k
> buff
>                     755896k actv,  142816k in_d,   15620k in_c
> Swap: 1999992k av,  324624k used, 1675368k free                  604408k
> cache
> 
> It can be a SCSI adapter issue ??

Could be. However, try to get accustomed to the "sysstat" rpm - it's a
really useful monitoring tool (among others, it provides both disk
subsystem statistics, virtual memory usage and CPU utilization). Have a
look at it.

It happened to my home system when it was used as: webserver (php+
mysql), mailserver AND... workstation! (Yeah, it had my wife's site on
it and it was the only system available at home. :(( ) All on 512 MB
RAM...

Sysstat proved that when being used as workstation, the system CPUs
(Intel HT) would spend between 20% and 75% in "idle" state due to huge
swapping. Doubling the memory to 1 GB solved the issue. (ok, it's not
serving HTTP and SMTP / IMAP nowadays, but doubling the RAM had a
tremendous impact on performance by then...)


Regards,
-- 
Andrei Pascal
RHCX
ITtraining.ro
Tel: +40741-200025
http://www.ittraining.ro