So it could be the apache server then? I do notice most of the programs run
pri between 15 and 20? Here is a snap of the top:
11:15pm up 14 days, 23:57, 2 users, load average: 1.34, 0.81, 0.76
70 processes: 63 sleeping, 7 running, 0 zombie, 0 stopped
CPU states: 3.0% user, 1.9% system, 0.0% nice, 2.2% idle
Mem: 257408K av, 170252K used, 87156K free, 44632K shrd, 39552K
buff
Swap: 265528K av, 1984K used, 263544K free 78700K
cached
PID USER PRI NI SIZE RSS SHARE STAT %CPU %MEM CTIME COMMAND
1442 nobody 20 0 2604 2604 1184 R 32.7 1.0 0:01 welcome.pl
1439 nobody 20 0 2448 2448 1220 R 23.1 0.9 0:07 who.pl
1441 nobody 10 0 2604 2604 1184 R 20.2 1.0 0:01 welcome.pl
1443 nobody 10 0 2884 2884 1180 R 15.4 1.1 0:01 profiles.pl
1440 nobody 10 0 2920 2920 1220 R 5.7 1.1 0:04 profiles.pl
1444 root 7 0 1004 1004 784 R 1.9 0.3 0:00 top
1 root 0 0 80 68 52 S 0.0 0.0 3219m init
2 root 0 0 0 0 0 SW 0.0 0.0 0:00 kflushd
3 root 0 0 0 0 0 SW 0.0 0.0 3:30 kupdate
4 root 0 0 0 0 0 SW 0.0 0.0 2:06 kswapd
5 root 0 0 0 0 0 SW 0.0 0.0 0:00 keventd
86 root 0 0 496 492 428 S 0.0 0.1 21:25 syslogd
89 root 0 0 480 0 0 SW 0.0 0.0 0:00 klogd
91 root 0 0 128 68 48 S 0.0 0.0 7:26 inetd
107 root 0 0 336 308 264 S 0.0 0.1 4:57 crond
109 daemon 0 0 128 56 48 S 0.0 0.0 0:00 atd
116 root 0 0 668 484 388 S 0.0 0.1 12:47 sendmail
Does this help any?
Thanks,
John
>
> I think you need to differentiate between the time your perl progs take to
> run, and the time the httpd daemon takes to complete the request. Very
> likely the httpd daemon puts your clients in a TIME_WAIT state. Do a
> "netstat -a" to see how many clients are put on hold. Let me put it this
> way: I would be seriously surprised when your perl progs would actually
take
> 3 minutes to complete. Although your load averages are quite staggering,
8%
> CPU per process is quite a lot.
>
> > I beleive this is due to the fact that say 15 people requesting
> > and running a perl proccess it splits the cpu between all 15
> > and each perl proccess on gets 8% or so give or take. The perl
> > proccess that normally only take a split second to run and display
> > at 100% when they are each using only 8 percent or so take
> > ungodly amount of time to run. How do you best attack
> > this issue??
>
> Well, the "ungodly" time they take to run, is actually for your own
> protection. :) I see these processes run at a "nice"value of 20 (very low
> priority). That might account for why they take so long to process. But it
> does not account for the 10%+ CPU usage per process. Normally, a "niced
20"
> process should have a negligible impact on your system--even 15 of them.
_______________________________________________
Perl-Win32-Web mailing list
[EMAIL PROTECTED]
http://listserv.ActiveState.com/mailman/listinfo/perl-win32-web
