Load avg issue

Andrei Pascal andrei at romsym.ro
Fri Apr 13 05:46:37 UTC 2007


On Fri, 2007-04-13 at 01:15 -0400, nilesh vaghela wrote:
> We have mail server running with sendmail, horde with front end and around
> 350 users.
> 
> Some time the server indicate very high cpu usage.
> 
> What is to be ckeck
> 
> We are using RHEL3.
> 
> 
>  10:49:11  up 91 days, 18:55,  3 users,  load average: 11.93, 15.27, 13.87
> 259 processes: 254 sleeping, 4 running, 1 zombie, 0 stopped
> CPU states:  cpu    user    nice  system    irq  softirq  iowait    idle
>            total   66.8%    0.0%   13.2%   0.4%     0.0%  118.8%    0.0%
>            cpu00   39.9%    0.0%    7.5%   0.5%     0.1%   51.6%    0.0%
>            cpu01   27.0%    0.0%    5.7%   0.0%     0.0%   67.1%    0.0%
> Mem:  1024780k av, 1008968k used,   15812k free,       0k shrd,   13764k
> buff
>                     749248k actv,  143504k in_d,   14280k in_c
> Swap: 1999992k av,  437476k used, 1562516k free                  558700k
> cached
> --------------
> [root at mailserver root]# ps -el | grep imap | wc -l
>     109
> -----------------

The "iowait" field in the CPU usage status means your processess are
waiting for some data. Check you disk and network latency. 1 GB RAM
should do it, but your swap is bieng hit quite hard. Check vmstat for si
and so (swap in / swap out) - I think that's where youp problem is.

Try stopping unnecessary services or simply add an extra GB of RAM to
that machine.


Regards,
Andrei





More information about the redhat-list mailing list