Re: [cobalt-users] Load avarage

["cbtrussell" <cbtrussell@xxxxxxxxxxx>]

> Normally my box has a load avarage of 0.00 to 0.08 or something for
> normal operation, and ofcourse some spikes to 0.20 when some cronjob
> runs.
>
> A few days ago it had to be rebooted, and now my load avarage is always
> around 0.20 to 0.30. It doesn't matter if it's busy or not, it's always
> around that figure. The reboot was because of a power failure, no extra
> stuff installed or updated.
>
> The box "does" around 3Gb a day of traffic, mostly web and some mail.
> No funny stuff running on it.
>
> What's a normal figure for this? And does anyone has an idea where this
> load may come from? When I check out top, I just see some httpd
> processes taking up 0.8% or something. It says 2.5% user time, 5%
> system time and 9x% idle.

I don't have a good explanation for you, but I've seen the EXACT same thing.
Particularly after a rough shutdown like a power failure. I don't know why
it does it, but I always wind up rebooting the machine through a normal
restart and the problem goes away. Usually, since the machine otherwise
appears to behave OK, I'll let it run with the higher load until after
midnight, then restart to minimize the disruption to my customers. Sometimes
it takes an additional restart or 2 (!!) to get things settled down. Of
course, you can't really know for a few hours after the reboot as your
RaQ(s) may do lots of little things after they boot that will impact the
load average.

OPEN QUESTION TO ALL: Does anyone know why this happens??

FYI, an excellent tool for monitoring load average is MRTG, it can be
configured to show you graphs for load average (among other things) over the
past 24 hrs, week, month, and year. Great tool for keeping tabs on past
performance.

http://people.ee.ethz.ch/~oetiker/webtools/mrtg/

Search the archives for "pointclark", you'll find a great rpm and script,
ready to go....

http://shack.pointclark.net/load.html

HTH,

Brandon