Very Computer

Hello,

I've been keeping track of machine statistics using 'sar' for the past
month.  I've mostly been paying attention to the runqueue sizes, and our
machine seems overloaded (runqueues of 6 on average, spikes up to 13;  I
understand a good target is around 2, with 4 being considered high).  In
any case, the statistics have remained consistant, up until last
friday.  An average day would look something like this:
---------------------
intserv1: sar -q

SunOS idev 5.5.1 Generic_103640-08 sun4u    04/09/98

12:06:00 runq-sz %runocc swpq-sz %swpocc
12:07:00     4.4     100                
12:08:00     5.5     100                
12:09:00     4.3     100                
12:10:00     3.2     100                
12:11:00     7.8      99                
12:12:00     5.7     101                
12:13:00     4.2     100                
12:14:00     6.9     100
---------------------

On friday, and since, however, it has been looking like this:

--------------------
intserv1: sar -q

SunOS idev 5.5.1 Generic_103640-08 sun4u    04/24/98

10:11:00 runq-sz %runocc swpq-sz %swpocc
10:12:00    16.7      10
10:13:00     2.0       7
10:14:00    18.4      15
10:15:00    11.9      15
10:16:00     3.8      20
10:17:00    13.6       8
10:18:00     6.0      10
10:19:00    17.4      13
---------------------

The runqueue size seems very erratic, and %runocc is much lower
(normally at %100).  Swap, memory, and everything else seem pretty much
ok though, as far as I can tell.

Can anyone shed some light on what might cause this?  There have not
been any significant changes made to the machine.  I get the feeling
that something is broken, either 'sar' or something else on the machine.

Also, I am running 'sar' every minute from cron, and at the same time
this started happening, I get blank entries in the log.  There is a time
stamp, just no data.

Any help would be appreciated.

Brian Mathis

PS> I'm also looking for a good place to find information of
interpreting 'sar' data.