On Mon, Dec 8, 2008 at 11:26, <[EMAIL PROTECTED]> wrote:
> Looking more closely at how find_avg() is used by cluster_view.php:
>
> $avg_cpu_num = find_avg($clustername, "", "cpu_num");
> if ($avg_cpu_num == 0) $avg_cpu_num = 1;
> $cluster_util = sprintf("%.0f", ((double) find_avg($clustername, "",
> "load_one") / $avg_cpu_num ) * 100);
>
> Basically, the code above takes the quotient of two averages:
>
> util = average(load_one) / average(cpu_num)
>
> Is that really the same as taking the average of the quotient?
>
> X = set of values for each t, (load_one(t) / cpu_num(t))
> util = average(X)
Computing the values on a "global" or "per-system" focus has wildly
different answers.
Consider two systems show on this URL:
http://spreadsheets.google.com/pub?key=pLqhPr4caFmQ3g_G7cwUCSQ
It shows a cluster makde of two dissimilar systems. One with 4 CPUs,
and a load of 4. One with one CPU, and a load of 0. Is your cluster
utilization 80% or 50%?
I (currently) believe that the current method, which is two divide the
average(load) by the average(cpu_num), is correct behavior. Computing
multiple per-system utilization figures, then averaging them, is less
immediately useful. (This is not to say "useless", since these
per-system utilization numbers could help to give you an idea what
rackspace is being efficiently used.)
--
Jesse Becker
GPG Fingerprint -- BD00 7AA4 4483 AFCC 82D0 2720 0083 0931 9A2B 06A2
------------------------------------------------------------------------------
SF.Net email is Sponsored by MIX09, March 18-20, 2009 in Las Vegas, Nevada.
The future of the web can't happen without you. Join us at MIX09 to help
pave the way to the Next Web now. Learn more and register at
http://ad.doubleclick.net/clk;208669438;13503038;i?http://2009.visitmix.com/
_______________________________________________
Ganglia-developers mailing list
Ganglia-developers@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-developers
