Hal,
This is running on a HAL (yes, it's seriously made by Hal, Hal :) ) SL100R,
pretty much an Sun Ultra 10 clone running w/ a 333mhz CPU and 256 megs of
ram. Memory doesn't seem to be an issue as there is only an occasional swap,
nor does IO seem to be an issue. I have some loggin scripts that pretty much
point to the CPU use being very high between 8am to 5pm. Below are some pretty
tyical outputs.
TOP
last pid: 5883; load averages: 1.14, 1.11, 1.15
15:19:58
45 processes: 41 sleeping, 1 running, 2 zombie, 1 on cpu
CPU states: 0.0% idle, 24.9% user, 75.1% kernel, 0.0% iowait, 0.0% swap
Memory: 256M real, 9288K free, 17M swap in use, 495M swap free
PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU COMMAND
5401 root 3 0 0 16M 8656K run 25:38 83.64% fw
14639 root 1 33 0 17M 14M sleep 67:28 10.85% ufpd
5873 root 1 58 0 2112K 1680K cpu 0:00 0.26% top
265 root 1 59 0 12M 8704K sleep 28:27 0.00% Xsun
1 root 1 48 0 688K 192K sleep 2:46 0.00% init
347 root 3 59 0 14M 4448K sleep 1:55 0.00% fw
353 root 3 58 0 14M 3816K sleep 0:39 0.00% fw
187 root 1 48 0 1592K 616K sleep 0:25 0.00% cron
2925 root 4 59 0 21M 16M sleep 0:07 0.00% fwui
351 root 3 59 0 59M 2264K sleep 0:04 0.00% fw
219 root 1 59 0 912K 448K sleep 0:02 0.00% utmpd
2986 root 1 39 10 3080K 1768K sleep 0:01 0.00% xlock
360 root 3 59 0 13M 832K sleep 0:01 0.00% fw
264 root 1 59 0 1824K 560K sleep 0:01 0.00% rpc.rstatd
159 root 4 11 0 2016K 888K sleep 0:00 0.00% keyserv
the "ufpd" process is our surf filtering application and is goes in and out of
sleep mode. I use top quite often on this box and this is a very typical
output during the day.
SAR:
06:00:00 1 1 0 98
06:05:00 1 1 0 98
06:10:00 1 1 0 98
06:15:00 1 1 0 98
06:20:00 1 1 0 98
06:25:00 2 2 0 96
06:30:00 2 2 0 95
06:35:00 6 9 0 84
06:40:00 3 2 0 94
06:45:00 3 2 0 95
06:50:00 2 2 0 96
06:55:00 3 2 0 95
07:00:00 40 59 0 0
07:05:00 3 2 0 95
07:10:00 12 19 0 69
07:15:00 5 8 0 87
07:20:00 6 10 0 84
07:25:00 25 68 0 7
07:30:00 9 20 0 71
07:35:00 5 5 0 90
07:40:00 9 15 0 76
07:45:00 14 28 0 58
07:50:00 10 16 0 74
07:55:00 16 24 0 60
08:00:01 25 72 0 3
08:05:01 11 17 0 71
08:10:01 15 17 0 68
08:15:01 15 21 0 65
08:20:01 20 46 0 34
08:25:01 30 70 0 0
08:30:01 29 64 0 7
08:35:01 29 65 0 6
08:40:01 26 57 0 17
08:45:01 29 71 0 0
08:50:01 31 69 0 0
08:55:01 22 54 0 24
09:00:00 19 23 0 59
09:05:00 24 56 0 20
09:10:00 27 62 0 11
09:15:00 31 68 0 1
09:20:00 28 72 0 0
09:25:00 26 68 0 5
09:30:00 20 46 0 34
09:35:00 24 43 0 32
09:40:00 29 69 0 2
09:45:00 28 72 0 0
09:50:00 25 63 0 13
09:55:00 30 70 0 0
10:00:00 27 73 0 0
10:05:00 31 69 0 0
10:10:00 29 71 0 0
10:15:00 30 70 0 0
10:20:00 28 72 0 0
10:25:00 29 71 0 0
10:30:00 30 70 0 0
10:35:00 31 69 0 0
10:40:00 30 64 0 6
10:45:00 30 70 0 0
10:50:00 31 69 0 0
10:55:00 30 70 0 0
11:00:00 31 69 0 0
11:05:00 30 70 0 0
11:10:00 25 45 0 29
11:15:00 32 68 0 0
11:20:00 30 60 0 9
11:25:00 32 52 0 16
11:30:00 30 61 0 9
11:35:00 22 49 0 29
11:40:00 24 51 0 26
11:45:00 28 72 0 0
11:50:00 26 74 0 0
11:55:00 30 70 0 0
12:00:01 30 58 0 12
12:05:01 20 38 0 42
12:10:01 28 67 0 5
12:15:02 24 52 0 24
12:20:02 31 69 0 0
12:25:02 31 69 0 0
12:30:02 28 72 0 0
12:35:02 27 73 0 0
12:40:02 31 69 0 0
12:45:02 32 68 0 0
12:50:02 33 67 0 0
12:55:02 32 65 0 4
13:00:00 32 68 0 0
13:05:00 30 70 0 0
13:10:00 35 64 0 1
13:15:00 30 69 0 1
13:20:00 31 64 0 5
13:25:00 31 64 0 5
13:30:00 31 65 0 5
13:35:00 30 70 0 0
13:40:00 32 68 0 0
13:45:00 32 68 0 0
13:50:00 28 62 0 10
13:55:00 30 70 0 0
14:00:00 27 56 0 18
14:05:00 29 71 0 0
14:10:00 21 45 0 34
14:15:00 28 48 0 24
14:20:00 33 67 0 0
14:25:00 29 71 0 0
14:30:00 29 71 0 0
14:35:00 31 69 0 0
14:40:00 32 66 0 2
14:45:00 31 66 0 3
14:50:00 33 60 0 7
14:55:00 30 70 0 0
15:00:00 31 69 0 0
15:05:00 30 70 0 0
15:10:00 29 71 0 0
15:15:00 25 58 0 17
15:20:00 28 71 0 0
This is the outpus since 6am, as you can see, utilization goes way up during the
day, and is practically idle during off hours
So what's the conclusion? basically need to throw more hardware at it?
Thanks,
Mike
On Mon, 31 Jul 2000, Hal Dorsman wrote:
> What kind of hardware? What utilization are you seeing?
> When utilization is high, what are your top processes?
> Do you have 'top'? If not, get it from sunsite.unc.edu.
> Are you running a security server (in.ahttpd)? I have intermittent
> problems with the security server that runs Websense running
> my utilization up very high, but then it drops down to
> normal (about 95% idle). This is supposedly a bug in in.ahttpd
> that I have never heard about a fix for. Doesn't seem to
> cause an performance issues so I have put that one on the low
> priority list.
>
> Hal
>
> Hal Dorsman
> Data Network Engineer
> Blackfoot Telephone Cooperative
> Missoula, Montana, USA
> [EMAIL PROTECTED]
> (406) 541-5106
>
>
> -----Original Message-----
> From: Michael Louie [mailto:[EMAIL PROTECTED]]
> Sent: Monday, July 31, 2000 1:34 PM
> To: [EMAIL PROTECTED]
> Subject: [FW1] Firewall CPU utilization
>
>
>
> Hi-
>
> ANyone have any ideas on what can be done to lower CPU load on a solaris 2.6
> box
> running version 4.0? I've already backed off my encryption to single DES
> and
> tried backing off o nthe logging, any other thoughts or ideas?
>
>
> Thanks,
> Mike
>
>
>
> ============================================================================
> ====
> To unsubscribe from this mailing list, please see the instructions at
> http://www.checkpoint.com/services/mailing.html
> ============================================================================
> ====
>
================================================================================
To unsubscribe from this mailing list, please see the instructions at
http://www.checkpoint.com/services/mailing.html
================================================================================
