[wellylug] Load average statistics dubious

[Atom Smasher]

On Tue, 30 Mar 2010, C.T.F. Jansen wrote:

> I used to use a number of different unix OS's and the load average 
> statistics on them is an artificial number and often with values too 
> different with what is actually happening on the system. It was regarded 
> as a joke and people warned not to use it.
=============

http://www.teamquest.com/resources/gunther/display/5/

it may not translate into mS of lag, and a load of (let's say) "3" on one 
system might not be a problem while that same load on another system might 
be terrible. but it's not a useless metric.

i'm not in any state to go through the maths in the link, above, but IIRC 
the load average depends on number of CPUs... so on a system with 32 CPUs, 
a load average under 32 (or is it 64?) shouldn't be a problem. that will 
further confuse things if you're just looking at the load average without 
proper context, and why my zshrc colors the load average based on how many 
CPUs it counts.

IIRC, the highest load average i've seen was >125, on a freeBSD box with 2 
CPUs. that was... impressive.


-- 
         ...atom

  ________________________
  http://atom.smasher.org/
  762A 3B98 A3C3 96C9 C6B7 582A B88D 52E4 D9F5 7808
  -------------------------------------------------

 	"Microsoft shouldn't be broken up. It should be shut down."
 		-- Bruce Schneier, 15 May 2000