Solaris

Hello Friends,
On one of my Solaris 10 box, CPU usage shows 100% using "sar", "vmstat". However, it has 4 CPUs and prstat and glance are not showing enough processes to justify high CPU utilization.

=========================================================================
$ prstat -a
PID USERNAME SIZE RSS STATE PRI NICE TIME CPU PROCESS/NLWP
1 root 2888K 1456K sleep 59 0 8:47:41 0.8% init/1
21678 yxhrprd 626M 527M sleep 59 0 5:35:08 0.5% java/47
27018 yxhrprd 38M 31M sleep 59 0 0:00:01 0.5% RTSDGN/1
18789 yxhrprd 1824K 1448K sleep 0 1 3:04:53 0.5% send2hotdir.ksh/1
19724 yxhrprd 1824K 1456K sleep 0 1 3:04:23 0.5% send2hotdir.ksh/1
24581 yxhrprd 1824K 1456K sleep 0 1 3:03:54 0.5% send2hotdir.ksh/1
18689 yxhrprd 1824K 1448K sleep 0 1 3:05:21 0.5% send2hotdir.ksh/1
27029 yxhrprd 515M 464M sleep 59 0 0:00:00 0.3% oracle/1
6299 yxecs 92M 84M sleep 59 0 32:30:56 0.3% dataserver/21
========================================================
# sar -u
00:00:00 %usr %sys %wio %idle
01:00:00 69 31 0 0
02:00:00 70 30 0 0
03:00:00 69 31 0 0
04:00:01 73 27 0 0
05:00:00 69 31 0 0
06:00:00 69 31 0 0
07:00:01 72 28 0 0
08:00:00 71 29 0 0
08:20:00 67 33 0 0
08:40:00 68 32 0 0
09:00:00 71 29 0 0
09:20:00 71 29 0 0
09:40:00 68 32 0 0
10:00:00 68 32 0 0
10:20:01 66 34 0 0
10:40:00 66 34 0 0
11:00:01 67 33 0 0
11:20:00 67 33 0 0
11:40:00 68 32 0 0
12:00:00 67 33 0 0
12:20:00 65 35 0 0
12:40:00 65 35 0 0
13:00:00 68 32 0 0
13:20:00 75 25 0 0
13:40:00 73 27 0 0
14:00:00 69 31 0 0
14:20:00 69 31 0 0
14:40:00 69 31 0 0
15:00:00 69 31 0 0
Average 69 31 0 0
========================================================

Any help on this would be highly appreciated.

Regards,
Mahi.

have u tried top in that machine. Whats the output?

Top is not installed on this system.

You may run :

Code:

ps -ef | cut -c42-100 | sort -nr | head

This will give you the time each process has consumed as percentage during its entire lifetime. 'cut' cuts off the unnecessary fields, thus putting focus on the TIME and CMD values, then sort them descending, and limit the output to the to ten consumers.

Looking at what you posted. I think the problem is most like a small set o resources being used by a large number of processes. It is not one process, but the sum of them all that is chewing cycles.