5 second cpu spike in FC3T3

[Maciej Żenczykowski]

you can change the top refresh rate (the 's' key), and try to catch it in 
the act :)

On Sun, 31 Oct 2004, W. Guy Thomas wrote:

> On Sun, 2004-10-31 at 21:33 +0100, Kyrre Ness Sjobak wrote:
> > what does top show?
> > 
> > s?n, 31.10.2004 kl. 05.08 skrev W. Guy Thomas:
> > > On Sun, 2004-10-31 at 00:04 -0400, Robert L Cochran wrote:
> > > > >
> > > > >
> > > > >>Could yoy please provide a bit more description?  What do you mean by a spike?
> > > > >>What kind of cpu and other hardware? Software versions of testing components 
> > > > >>(if upgraded)?
> > > > >>
> > > > >>Tom
> > > > >>-- 
> > > > >>Tom Taylor 
> > > > >>Registered linux user #263467
> > > > >>    
> > > > >>
> > > > >
> > > > >ok, fire up gkrellm.
> > > > >watch "cpu"
> > > > >every 5 seconds I get a complete cpu spike to the top of the meter.
> > > > >when pc goes into screensaver mode, every 5 seconds, xscreensaver halts
> > > > >for 1 second, then continues.
> > > > >
> > > > >I am on a Dell C840 laptop, PIII700, 384RAM, 40GB, nvidia graphics (from
> > > > >FC3T3 not nvidia)...nothing special as far as testing, this is stock
> > > > >FC3T3.
> > > > >
> > > > >Thanks.
> > > > >  
> > > > >
> > > > You might want to check the GKrellM web site first, the FAQ may be helpful.
> > > > 
> > > > http://web.wt.net/~billw/gkrellm/gkrellm.html
> > > > 
> > > > 
> > > 
> > > yes, I've looked at that.
> > > this didn't occur on FC1 or FC2 so I don't think it's gkrellm specific.
> > > I still think it's picking up some logging on ext3 or something like
> > > that.
> > > 
> > > 
> > > -- 
> > > =Guy
> > > 23:07:27 up 2 days, 1:36, 3 users, load average: 1.12, 1.24, 1.14
> > 
> 
> I don't see anything unusual in top. The spike is so quick, it's
> literally a straight up line in gkrellm, just every 5 seconds.
> 
> I'm not sure what else I can use to monitor this. I wouldn't even know
> about it if gkrellm hadn't shown it to me. It doesn't affect performance
> really, but it's annoying to know it's there.
> 
> any other ideas are welcome.
> 
>