Thanks guys for the thoughts.

The server does Nagios for us. The system does log everything, but the
iowait time is very low and hte number of iops is low. It's monitoring
>450 services, and can sometimes be pretty busy with little tiny jobs
like pinging a host 10 times, and such. Sometimes this makes the load
average high for a very short time. The CPU is not at 100% at these
times and the server is not feeling sluggish. It just has a lot of
things running. There is no reason for it to delay incoming connections.

As we have grown we have split apps up that were once together on the
same server. This box was once Cacti, Nagios and ticketing. Now it's
just Nagios and ticketing. We did not want any email to be delayed,
period. I just want to figure out how to adjust that load average that
kicks off those rejection of connections.... until the new servers are
ready for Nagios and the ticketing system to each have their own.

So, how do I adjust the number ?

Thanks,
Mike



On Mon, 2008-02-04 at 15:02 -0800, James G. Sack (jim) wrote:
> Neil Schneider wrote:
> > If it's only momentary and seldom, it shouldn't matter much. Servers should
> > resend mail to the server when they get this temporary refusal to connect. I
> > may be remembering wrong but I thought Postfix, and maybe sendmail would 
> > just
> > throttle back until the load average dropped and pass a service temporarily
> > unavailable message too.
> 
> IANA email-admin but..
> 
> Now that I think of it, I remember seeing runaway error message logging
> being the cause of such elevated loadavg values.
> 
> See if something is spamming your logs.
> 
> Regards,
> ..jim
> 
> 
-- 
************************************************************
Michael J. McCafferty
Principal, Security Engineer
M5 Hosting
http://www.m5hosting.com

You can have your own custom Dedicated Server up and running today !
RedHat Enterprise, CentOS, Fedora, Debian, OpenBSD, FreeBSD, and more
************************************************************


-- 
[email protected]
http://www.kernel-panic.org/cgi-bin/mailman/listinfo/kplug-list

Reply via email to