I've been getting e-mails from one of our raq4s' since about 3:30 am yesterday saying the loda average is to high and sendmail don't like it. Here is the output from top:


9:16am up 13:01, 4 users, load average: 7.55, 20.74, 42.86
100 processes: 92 sleeping, 8 running, 0 zombie, 0 stopped
CPU states: 54.0% user, 45.9% system, 0.0% nice, 0.0% idle
Mem: 257636K av, 182212K used, 75424K free, 344752K shrd, 7576K buff
Swap: 131532K av, 18056K used, 113476K free 78484K cached

PID USER PRI NI SIZE RSS SHARE STAT LIB %CPU %MEM TIME COMMAND
15914 root 16 0 708 680 340 R 0 20.9 0.2 12:05 top
15782 root 16 0 708 680 340 R 0 19.9 0.2 12:45 top
19941 httpd 5 0 10780 10M 8392 S 0 4.8 4.1 0:05 httpd
19954 httpd 6 0 10396 10M 8908 S 0 3.2 4.0 0:08 httpd
19953 httpd 7 0 10588 10M 8892 S 0 3.1 4.1 0:07 httpd
19939 httpd 7 0 10856 10M 8380 S 0 2.9 4.2 0:08 httpd
19951 httpd 0 0 10424 10M 8392 S 0 2.9 4.0 0:08 httpd
20485 httpd 0 0 10252 10M 8924 S 0 2.9 3.9 0:03 httpd
20572 root 5 0 916 916 680 R 0 2.9 0.3 0:04 top
19912 httpd 0 0 10248 10M 8904 S 0 2.5 3.9 0:07 httpd
19914 httpd 6 0 10420 10M 8908 S 0 2.5 4.0 0:06 httpd
19947 httpd 6 0 10416 10M 8892 S 0 2.3 4.0 0:07 httpd
20491 httpd 0 0 10256 10M 8924 S 0 2.3 3.9 0:02 httpd
20489 httpd 3 0 10252 10M 8928 S 0 2.1 3.9 0:01 httpd
20668 httpd 6 0 10348 10M 8924 S 0 2.1 4.0 0:01 httpd
20484 httpd 5 0 10232 9M 8912 S 0 1.9 3.9 0:02 httpd
20812 root 13 0 1832 1832 1352 R 0 0.7 0.7 0:00 sendmail
:eek:

Notice the RSS size of a lot of the httpd processes is 10 megs. Does this mean that each of those httpd processes is downloading 10 megs? If not, then could someone explain to me what might be happening to cause that 10M RSS and the high load average.

Thanks in advance.:erm:

Part of the problem is there is 3 top processes running. top is resource hog. Kill them off.

2nd problem is the system is using 45% of the resources to handle the load - not good numbers.

Yes. each of the httpd procs is using 10MB. I'd take a look at any new cgi or mail stuff thats going on. Any chance someone is hijacking formail.pl?

Anyway, restart apache and see what happens. If it still doesn't free up memory, reboot and watch it closely.

Tom

jucebro,

Did you find out what was causing the 10M httpd processes?

yeah please also share it when you solve a problem. We can learn from eachother :confused:

Originally posted by tom.oneil

Yes. each of the httpd procs is using 10MB. I'd take a look at any new cgi or mail stuff thats going on. Any chance someone is hijacking formail.pl?
Tom

Actually, each httpd process is *NOT* using 10 megs each. The RSS is not a cumulative amount for parent processes with children. Many of those httpd processes share the same memory space, so this is not a 100% valid statement.

I run Chilisoft ASP, and the RSS is 40 megs, times 15 children, would be 600 megs, which is more than the machine has. And I have 300 megs free at the moment :)

- John C.