[NTLUG:Discuss] Top question

[Patrick R. Michaud]

On Wed, Jan 21, 2004 at 06:49:50PM -0600, Greg Edwards wrote:
> I've been trying to find out exactly what the load averages displayed by 
> top really mean.  

>From the man page for "top" (RH9):

   The load averages are the average number of processes ready to run
   during the last 1, 5, and 15 minutes.

What this means in english:

On a multitasking system such as Linux, there can be many processes
running concurrently.  While one process is running (using the processor),
there may be other processes that are ready to run as soon as the processor
becomes available or is otherwise switched to another process.  The 1-minute
load average is the average number of processes that were in this "waiting
for the processor" state during the last 1 minute. 

Higher numbers generally indicate more heavily loaded systems.  In my
experience load averages less than 5.00 are pretty normal, load averages
between 5.00 and 10.00 mean the system is somewhat heavily loaded (and
it might be worth finding out what's up), and load averages more than 10.00
usually mean something is seriously loading the system and it's definitely
worth looking into the causes.  (Other people may have other thresholds
for load averages--these are just the ones I've found typical.)

And for server admins: note that some programs such as sendmail begin
to refuse incoming connections when the load average gets high, in 
order to reduce the load and wait to process incoming mail when the
system is more lightly loaded.  

Pm
--------------
Patrick R. Michaud, RHCE #808002519807115
Web:   http://www.pmichaud.com
Email: pmichaud at pobox.com