FW: Strange HI Load

[Glenn Steen]

On 17/06/06, Thomas Chamtieh <tchamtieh at nayzak.com> wrote:
> Here's an output of 'ps aux' look at all these MailScanner
> processes!!!!:
>
(snip)
>
>
> A 'top' will give the following, I noticed that it's starting to use
> swap at this point:
>
>  17:04:48  up 7 days, 17:46,  1 user,  load average: 3.66, 2.00, 1.38
> 271 processes: 269 sleeping, 2 running, 0 zombie, 0 stopped
> CPU states:  cpu    user    nice  system    irq  softirq  iowait    idle
>            total   43.8%    0.0%    7.3%   0.0%     0.0%   48.7%    0.0%
>            cpu00   31.3%    0.0%    5.8%   0.0%     0.0%   62.7%    0.0%
>            cpu01   45.0%    0.0%   13.7%   0.0%     0.0%   41.1%    0.0%
>            cpu02   37.2%    0.0%    1.9%   0.0%     0.0%   60.7%    0.0%
>            cpu03   61.7%    0.0%    7.8%   0.0%     0.0%   30.3%    0.0%
> Mem:  3082456k av, 3017552k used,   64904k free,       0k shrd,   24568k
> buff
>                    2327328k actv,  431144k in_d,   39764k in_c
> Swap: 4194232k av,  620560k used, 3573672k free                  236836k
> cached
>
>   PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME CPU
> COMMAND
> 13862 root      22   0 38828  30M  2528 R    16.9  1.0   0:01   3
> MailScanner
> 13785 root      16   0 45444  39M  2568 S     2.9  1.3   0:00   2
> MailScanner
> 13777 root      15   0 44532  25M  2340 D     1.9  0.8   0:00   2
> MailScanner

You have a 4 CPU machine with a load under 4 == You don't have quite a
process runnable/processor. Not the problem, nor any real indicator.

That it uses swap *might* be an indicator. Do the usual "vmstat 2" and
look if there seem to be swap activity most all of the time... That
could really hurt performance.

When you stop MailScanner manually, does all the children (eventually) die?

-- 
-- Glenn
email: glenn < dot > steen < at > gmail < dot > com
work: glenn < dot > steen < at > ap1 < dot > se