Apache CPU usage explodes since disastrous migration to 10.0.1

[regisg]

During last day :

Durint last 3 hours :

Server health parameter "Services > Apache CPU usage" changed its status from "yellow" to "red".
top - 19:44:05 up 2 days, 6:31, 0 users, load average: 1.34, 0.95, 0.45
Tasks: 56 total, 3 running, 53 sleeping, 0 stopped, 0 zombie
Cpu(s): 3.1% us, 1.3% sy, 0.1% ni, 95.4% id, 0.1% wa, 0.0% hi, 0.0% si
Mem: 2097152k total, 730112k used, 1367040k free, 0k buffers
Swap: 0k total, 0k used, 0k free, 0k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
14198 apache 25 0 326m 39m 5312 R 99.2 1.9 3:40.83 httpd
3340 apache 21 0 319m 35m 5884 R 99.2 1.7 0:18.83 httpd

Server health parameter "Services > Apache memory usage" changed its status from "green" to "yellow".
top - 22:44:17 up 2 days, 9:32, 0 users, load average: 9.11, 9.30, 9.34
Tasks: 60 total, 10 running, 50 sleeping, 0 stopped, 0 zombie
Cpu(s): 3.8% us, 1.8% sy, 0.1% ni, 94.2% id, 0.1% wa, 0.0% hi, 0.0% si
Mem: 2097152k total, 1035228k used, 1061924k free, 0k buffers
Swap: 0k total, 0k used, 0k free, 0k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
28431 apache 25 0 342m 51m 5980 R 27.6 2.5 27:27.12 httpd
25616 apache 25 0 335m 46m 4448 R 27.6 2.3 19:41.55 httpd
10235 apache 25 0 348m 64m 5236 R 25.7 3.1 29:01.89 httpd
11777 apache 25 0 340m 55m 5196 R 25.7 2.7 43:21.69 httpd
15486 apache 25 0 344m 60m 5344 R 25.7 3.0 16:53.33 httpd
20324 apache 25 0 337m 51m 4772 R 25.7 2.5 22:10.77 httpd
15476 apache 25 0 340m 55m 3960 R 23.7 2.7 37:14.76 httpd
26112 apache 25 0 326m 41m 5280 R 17.8 2.0 7:24.88 httpd
24283 apache 25 0 327m 39m 5260 R 5.9 1.9 9:51.54 httpd

Thus, very often, apache stops itself...

(Apache httpd.conf parameters are all setting to defauts values)

Any help would be very appreciated !

Thx a lot.

 

[SandorB]

Hi Regisg,

I had the same until I set

PHP support (run as FASTCGI application , PHP 'safe_mode' on)

for all domains.

I hope it will help.

Best wishes

Sándor

 

[regisg]

Thx Sandor, i will try...

Anyway, it looks contradictory :

http://download1.parallels.com/Ples...plesk-control-panel-user-guide/toc6461316.htm

# Apache module (only on Unix). This mode uses less server resources, but it is the least secure option. A site is isolated only if PHP safe mode is switched on.

# FastCGI application. This mode is faster than CGI, but consumes more memory. Provides site isolation.

 

[SandorB]

Hi Regisg,

I know that it looks contradictory but I tried all the variations of the settings. I also read the manual but for me there is no more CPU overload. I also know that it should work smoothly with all the variations of the settings but for me it does not. This settings is safe and FOR ME solved the problem.

I hope it will help you as well.

Best wishes

Sándor

 

[xmuskrat]

Changing from FastCGI to Apache modules fixed the issue for me.

 

[regisg]

So that's the opposite of SandorB... !
I'am not sure this is a serious lead.

I think the problem is linked to watchdog, see http://forum.parallels.com/showthread.php?t=106130

I notices that i have also the same error :
Cron <root@munci> /usr/local/psa/libexec/modules/watchdog/cp/send-report weekly
PHP Fatal error: Class 'OptionalServices_ServicesConfiguration' not found in /usr/local/psa/admin/plib/common_func.php3 on line 2948

 

[DWMTS1]

Did anyone find resolution to this? I have the same issue