Strange perl processes executed by user Apache

[goldbar]

Dear All! I have been getting strange issues with one of my D.A servers running on CentOS X86_64.
Every now and then system load averages shoot upto 8.00 and 9.00.
Investigations show processes like these
14395 apache 16 0 69332 6652 1920 S 0 0.2 21:16.83 perl
15302 apache 16 0 69332 6652 1920 S 0 0.2 21:07.33 perl
11024 apache 16 0 69332 6648 1920 S 0 0.2 21:32.98 perl
14168 apache 15 0 69332 6648 1920 S 0 0.2 21:00.53 perl
14258 apache 16 0 69332 6648 1920 S 0 0.2 23:09.43 perl
9294 apache 16 0 69332 6644 1920 S 0 0.2 63:41.51 perl
13960 apache 16 0 69332 6644 1920 S 0 0.2 22:23.37 perl
5580 root 16 0 19428 6548 1260 S 0 0.2 0:19.60 hald
21341 apache 16 0 69588 6248 1308 S 0 0.2 0:00.66 perl
21350 apache 16 0 69588 6248 1308 S 0 0.2 0:00.62 perl
11025 apache 16 0 69332 6112 1384 S 0 0.2 22:56.11 perl
14169 apache 16 0 69332 6112 1380 S 0 0.2 22:34.66 perl
14259 apache 16 0 69332 6112 1380 S 0 0.2 22:29.10 perl
14396 apache 16 0 69332 6112 1380 S 0 0.2 21:18.95 perl
15303 apache 16 0 69332 6112 1380 S 0 0.2 20:46.50 perl
9295 apache 16 0 69332 6108 1384 S 0 0.2 30:52.95 perl
13961 apache 15 0 69332 6108 1380 S 0 0.2 23:57.49 perl
16396 apache 16 0 69332 6108 1384 S 0 0.2 288:03.45 perl
Doing a ps auxf reveal httpd process eating up all the cpu. Sample ps auxf is like apache 16396 9.3 0.1 69332 6108 ? S Oct09 288:03 /usr/sbin/httpd
apache 9294 3.2 0.1 69332 6644 ? S Oct10 63:41 /usr/sbin/httpd


I have tried to figure out the scripts causing this via extended status etc but to no use.
Can anyone tell me how can i investigate as to what could be the reason of this.
Any help will be much appreciated.
Regards
Goldbar Hosting

 

[brinkie]

I have the same problem...!

Server configuration:
# ASUS M2A-VM (AMD)
# AMD Athlon(tm) 64 X2 Dual Core
# 4 GB (4x1) DDR2 667Mhz Kingston memory
# FreeBSD 6.2-Stable
# Apache 1.3.39
# Exim 4.62
# MySQL 5.0.22
# Perl 5.8.8

Now, what happens is that a perl-process is freaking out. When i run a (k)trace on the process id, i only get the information that Perl is doing a CALL, thousands of lines/second. CPU1 is beeing used 99-100%, CPU0 doesn't seem to be effected. So a trace doesn't give me much information.

This started about a week ago, and ever since i've been trying to find the website/user that is causing it.

I would be very gratefull if someone could point me to a simple tool that allows me to see who (what website/user/script) is causing this.

Ps. this is my first FreeBSD box, before i used linux on my machines... i'm lost, please help! I'm afraid that one day soon i won't be there on time to kill the process and it will run multiple times and will crash the machine (using both cores).

Pps. i tried to install PRM but untill now i've failed, PRM doesn't seem to work on this setup <sigh>..