high load

[laddu]

hello,

I am encountering some high load on my server:

Code:

top - 13:38:03 up 1 day, 53 min,  2 users,  load average: 30.79, 20.56, 15.28
Tasks: 280 total,   1 running, 279 sleeping,   0 stopped,   0 zombie
Cpu(s):  0.7%us,  0.8%sy,  0.0%ni,  4.8%id, 93.6%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   2074936k total,  2024312k used,    50624k free,     1392k buffers
Swap:  2031608k total,  1102600k used,   929008k free,    18204k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
8438 vu2005    17   0 24852 7612 5100 S    1  0.4   0:00.09 php-cgi
8403 vu2005    18   0 24692 6632 5024 D    1  0.3   0:00.11 php-cgi
8333 root      18   0 15164 2496 1520 D    0  0.1   0:00.32 ispcp-apache-lo
8367 apache    18   0 12812 2044 1212 D    0  0.1   0:00.03 httpd
8386 vu2003    18   0 26360 1704 1268 D    0  0.1   0:00.13 php-cgi
8423 vu2003    17   0 24692 6860 5012 D    0  0.3   0:00.15 php-cgi
8437 vu2003    17   0 23728 4640 3476 D    0  0.2   0:00.05 php-cgi
8450 vu2003    18   0 24852 7544 5076 S    0  0.4   0:00.09 php-cgi

Found some similar threads where the solution was to upgrade php. I am currently with the latest php version as you can see:

Code:

php -v
PHP 5.2.6 (cli) (built: May  7 2008 00:50:43)
Copyright (c) 1997-2008 The PHP Group
Zend Engine v2.2.0, Copyright (c) 1998-2008 Zend Technologies

I can see that php-cgi is causing the issue. Any idea about I can resolve this ?

Regards

[BeNe]

Which OS do you use ?
Is FastCGI or fcgid running ?

Greez BeNe

[laddu]

CentOS release 5.2 (Final)

apache 19258 0.0 0.1 12292 3272 ? S 14:27 0:00 /usr/sbin/fcgi-

I think fcgi is running.

[nuke3d]

Well, seems like you're having problems with the HDD. All the CPU time is spent on IO wait, that's not normal. Is it constantly like that?

[laddu]

Sorry

FastCGI was waorking. I now replaced the line with fcgid module.

I will update this with my result

---

(02-05-2009 11:00 PM)nuke3d Wrote:  Well, seems like you're having problems with the HDD. All the CPU time is spent on IO wait, that's not normal. Is it constantly like that?

It is working in some kind of loops. the memory goes down to 35mb and all the sites are returning 500 internal error, and after that it gets free a little bit, and so on

Now I am with fcgid instead of fastcgi.

[laddu]

Also found this:

Quote:Feb 5 16:41:16 ns1 ispcp_daemon[1530]: child 1530 started !
Feb 5 16:41:16 ns1 ispcp_daemon[1530]: Aeee! SIG_PIPE was received ! Will we survive ?
Feb 5 16:41:16 ns1 ispcp_daemon[1530]: send_line(): socket write error: Broken pipe
Feb 5 16:41:16 ns1 ispcp_daemon[1530]: read_line(): socket EOF ! other end closed the connection !
Feb 5 16:41:16 ns1 ispcp_daemon[3158]: EINTR was received ! continue;

I finished testing HDD with smartctl and all passed

Quote:[root@ns1 ~]# smartctl -d ata -H /dev/sda1
smartctl version 5.36 [i686-redhat-linux-gnu] Copyright © 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED
[root@ns1 ~]# smartctl -d ata -H /dev/sdb1
smartctl version 5.36 [i686-redhat-linux-gnu] Copyright © 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

[root@ns1 ~]# smartctl -d ata -H /dev/sdc1
smartctl version 5.36 [i686-redhat-linux-gnu] Copyright © 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

[root@ns1 ~]# smartctl -d ata -H /dev/sdd1
smartctl version 5.36 [i686-redhat-linux-gnu] Copyright © 2002-6 Bruce Allen
Home page is http://smartmontools.sourceforge.net/

=== START OF READ SMART DATA SECTION ===
SMART overall-health self-assessment test result: PASSED

Any better tool ? Any suggestion ?

[ephigenie]

It doesn't mean your hdd is broken.

Quote:Cpu(s): 0.7%us, 0.8%sy, 0.0%ni, 4.8%id, 93.6%wa, 0.0%hi, 0.0%si, 0.0%st

the 93,6% WA just means that your processor is waiting 93,6% of its time for its harddrive to handle over requested data.
It could be that there's some IO intensive third party process going on, using less CPU cycles but produces a lot of IO requests. At some point its just too much for your harddrive - and it cannot handle that load - thats why the whole server load gets high.

Because nearly everything will slow down then ...

[laddu]

Do you know any way to find witch process is creating so much wa ?

[KeNt]

(02-06-2009 01:18 AM)laddu Wrote:  Do you know any way to find witch process is creating so much wa ?

# top
Shift + p

[nuke3d]

1.1GB swap used... Do you have a process that eats all your ram? Probably your system is swapping constantly.