I did try the snap shot 4/14/2001 4.06dev, it didnt' solve the problem.
If 256mb ram , dual 400mhz can only serve around 10 concurrent clients, this
is a MAJOR issue for PHP. I guess any web language like PERL, Cold fusion,
JSP, servlet can do a LOT more than this with the equipment I am using right
now.
-----Original Message-----
From: Cameron
To: [EMAIL PROTECTED]
Sent: 4/14/01 12:45 PM
Subject: PHP 4.0 Bug #10299 Updated: CPU and Memory Spike
my suggestions
1. update to latest php snap from http://snaps.php.net/
they are "beta" but generally run better than the current release.
2. set your max execution time
change it to as close to possible to the max time any script would
really take, generally something around 3 seconds
3. more ram
256meg is good for upto about 10 concurrent clients, much more and the
server starts to grind, i have 192meg in a box that serves
5000pages/day.
4. try some caching?
http://apc.communityconnect.com/
try it out, it helped me a bit with making a server go a bit faster,
dont use the SHM version with that little ram tho, use the MMAP version.
Cameron Brunner
Senior Web Programmer
inetsalestech.com
[EMAIL PROTECTED] wrote:
> ID: 10299
> User Update by: [EMAIL PROTECTED]
> Status: Open
> Bug Type: Performance problem
> Description: CPU and Memory Spike
>
> This is vmstat show if I keep reloading the script
>
> procs memory page disks faults
cpu
> r b w avm fre flt re pi po fr sr da0 da1 in sy cs us
sy id
> 1 4 0 73860 13432 4 0 0 0 0 0 2 0 382 3804 258 16
1 83
> 0 4 0 73860 13424 4 0 0 0 0 0 0 0 357 3789 245 15
4 81
> 1 4 0 73860 13424 4 0 0 0 0 0 0 0 373 3763 263 14
3 83
> 2 4 0 73860 13416 142 0 0 0 139 0 0 0 402 6155 285 26
4 70
> 0 4 0 73860 13408 4 0 0 0 4 0 11 0 366 3179 220 11
2 87
> 1 4 0 73860 13400 4 0 0 0 0 0 0 0 358 3943 257 15
3 82
> 2 4 0 73860 13400 4 0 0 0 0 0 0 0 340 4201 240 12
42 45
> 2 4 0 73860 13400 4 0 0 0 0 0 0 0 240 1842 90 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 19 0 265 1802 98 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 241 1753 93 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 239 1729 96 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 241 1612 94 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 243 1576 97 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 243 1541 98 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 239 1487 102 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 241 1401 93 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 251 1405 100 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 246 1338 101 0
100 0
> 2 4 0 73116 13400 4 0 0 0 0 0 0 0 238 1297 94 0
100 0
> 2 4 0 73860 13400 4 0 0 0 0 0 0 2 243 1282 100 0
100 0
> 2 4 0 73860 13400 4 0 0 0 0 0 0 0 239 1223 89 0
100 0
> 2 4 0 73860 13400 4 0 0 0 2 0 9 16 300 1180 92 0
100 0
> procs memory page disks faults
cpu
> r b w avm fre flt re pi po fr sr da0 da1 in sy cs us
sy id
> 3 4 0 73860 13400 4 0 0 0 0 0 0 0 251 1156 90 0
100 0
> 3 4 0 50296 13400 4 0 0 0 0 0 6 0 258 1116 93 0
100 0
> 2 4 0 50296 13400 4 0 0 0 0 0 0 0 240 1108 83 0
100 0
> 3 4 0 50296 13400 4 0 0 0 0 0 0 0 245 1064 92 0
100 0
> 3 4 0 46988 13400 4 0 0 0 0 0 0 0 242 1031 83 0
100 0
> 3 4 0 46988 13400 4 0 0 0 0 0 0 0 242 989 82 0
100 0
> 2 4 0 46988 13400 4 0 0 0 0 0 0 0 243 1010 85 0
100 0
> 2 4 0 46988 13400 4 0 0 0 0 0 1 0 241 985 81 0
100 0
> 2 4 0 46988 13400 4 0 0 0 0 0 0 0 241 971 89 1
99 0
> 2 4 0 46988 13400 4 0 0 0 0 0 0 0 243 940 74 0
100 0
> 3 4 0 46988 13400 4 0 0 0 2 0 8 0 250 926 80 0
100 0
> 2 4 0 46988 13400 4 0 0 0 0 0 1 0 245 930 88 0
100 0
> 3 4 0 46988 13400 4 0 0 0 0 0 0 0 240 905 80 0
100 0
> 2 4 0 46988 13400 4 0 0 0 0 0 7 0 253 920 100 0
98 2
> 3 4 0 47332 13056 90 0 0 0 0 0 0 0 239 2113 116 3
97 0
> 2 4 0 46588 13048 4 0 0 0 0 0 1 0 243 2334 103 1
99 0
> 1 4 0 46244 13392 4 0 0 0 86 0 8 0 276 4383 140 7
19 74
> 1 4 0 47896 13184 59 0 0 0 0 0 2 0 275 621 56 7
1 92
> 0 4 0 47896 13152 20 0 0 0 0 0 1 0 253 471 46 6
1 93
> 1 4 0 47492 13372 18 0 0 0 64 0 0 0 251 478 43 0
1 99
> 0 4 0 47492 13372 4 0 0 0 0 0 0 0 235 440 40 0
0 100
> 0 4 0 47492 13372 4 0 0 0 0 0 0 0 238 442 40 0
2 98
>
> The CPU time drop from 80 all the way to zero and make the Apache hang
for like 30 seconds and the resume back again. It happens VERY often if
I only have like 30 users surf the website at the same time. It pretty
much bring the web server down if I have around 50 users surft the site
>
> Previous Comments:
>
------------------------------------------------------------------------
---
>
> [2001-04-11 18:03:59] [EMAIL PROTECTED]
> I am running on FreeBSD 4.2 Stable, Apache 1.3.19 + PHP 4.0.4pl1 +
Modssl on a Dual P2 400mhz with 256mb ram. From time to time, one of
the httpd process will eat up all the CPU time and memory, and after
around 5 to 10 seconds, it goes back to normal Even a Very simple
script like
>
> <?
> echo "test";
> ?>
>
> can make it happen. 1 out of 25 reload of the above script can make
the spike happen. But the web server do not crash though. If I run on
a script with heavy database access, it happen 1 out of 10 if I press
reload button continously.
>
> BTW, is that normal for Apache to take up 30% of the CPU time to
excute one single PHP-database-driven page?
>
> I search the bugs database and I found another user have the exact
problem like what I am having right now. His bug id is #9154.
>
> This is how I configure PHP
> ./configure --with-mysql=/usr/local/mysql
--with-apxs=/usr/local/apache/bin/apxs
--with-config-file-path=/usr/local/etc --enable-track-vars --enable-ftp
--with-gd=/usr/local --with-pcre-regex=/usr/local/
> --with-jpeg-dir=/usr/local --with-xpm-dir=/usr/X11R6
>
> With or Without modssl enable does not make a difference. Both can
cause the spike.
>
>
------------------------------------------------------------------------
---
>
> Full Bug description available at: http://bugs.php.net/?id=10299
>
> --
> PHP Development Mailing List <http://www.php.net/>
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, e-mail: [EMAIL PROTECTED]
> To contact the list administrators, e-mail:
[EMAIL PROTECTED]
--
PHP Development Mailing List <http://www.php.net/>
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
To contact the list administrators, e-mail: [EMAIL PROTECTED]
