Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

Recently moved to a new server with CentOS 5.4 and is 64 bit as opposed to our old 32-bit. It's got dual processors now and a bit higher memory.

I realized though, that apache has been heavily eating away at the memory and sometimes it would cause the system to swap.

The current setup has Nginx 1.0.8 in front to serve static content and Apache 2.2 to serve dynamic content (PHP). I also have APC installed. What could be causing this memory usage spike? I tried tuning the config file for apache but it would work if I turn certain directives to a absurdly low number. Before on the old machine I was able to use the default settings with no more than ~260MB of memory usage for all apache processes:

<IfModule prefork.c>
StartServers       8
MinSpareServers    5
MaxSpareServers   20
ServerLimit      256
MaxClients       256
MaxRequestsPerChild  4000
</IfModule>

But now the new machine is using this with high memory usage, it can get as high has 600MB+:

<IfModule prefork.c>
StartServers       5
MinSpareServers    5
MaxSpareServers   10
MaxClients        20
MaxRequestsPerChild  500
</IfModule>

APC settings are as follows:

extension = apc.so
apc.enabled=1
apc.shm_segments=1
apc.optimization=0
apc.shm_size=64M
apc.ttl=0
apc.user_ttl=7200
apc.num_files_hint=1024
apc.mmap_file_mask=/tmp/apc.XXXXXX
apc.enable_cli=1
apc.cache_by_default=1

Any ideas why this is happening?

Edit: Some more process info

top - 21:18:45 up 10 days,  6:20,  4 users,  load average: 0.00, 0.04, 0.06
Tasks: 103 total,   2 running, 101 sleeping,   0 stopped,   0 zombie
Cpu(s):  0.3%us,  0.0%sy,  0.0%ni, 98.8%id,  0.2%wa,  0.0%hi,  0.7%si,  0.0%st
Mem:   1541280k total,  1479804k used,    61476k free,    99812k buffers
Swap: 10482404k total,    25548k used, 10456856k free,   725180k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND                                                                                                                                     
 5613 jenkins   17   0 1144m 184m  13m S  0.7 12.3   2:05.94 java                                                                                                                                         
25306 apache    15   0  541m 125m  69m S  0.0  8.3   0:10.44 httpd                                                                                                                                        
25308 apache    15   0  533m 118m  67m S  0.0  7.9   0:11.01 httpd                                                                                                                                        
25356 apache    16   0  524m 112m  69m S  0.0  7.5   0:08.84 httpd                                                                                                                                        
25305 apache    15   0  510m 103m  69m S  0.0  6.9   0:11.58 httpd                                                                                                                                        
25309 apache    16   0  498m  92m  69m S  0.0  6.1   0:10.35 httpd                                                                                                                                        
25307 apache    15   0  499m  90m  67m S  0.0  6.0   0:09.59 httpd                                                                                                                                        
 9759 mysql     15   0  428m  83m 6920 S  0.0  5.6  11:10.01 mysqld                                                                                                                                       
16121 apache    15   0  494m  68m  48m S  0.0  4.6   0:05.50 httpd                                                                                                                                        
 2945 user      15   0  659m  18m 6364 S  0.0  1.2   0:00.53 node                                                                                                                                         
25303 root      15   0  425m  12m 6176 S  0.0  0.8   0:00.07 httpd                                                                                                                                        
 2944 user      15   0  636m  10m 2288 S  0.0  0.7   0:00.11 node  



[user@server~]# ps o rss,command -u apache
  RSS COMMAND
70236 /usr/sbin/httpd
105860 /usr/sbin/httpd
128096 /usr/sbin/httpd
92888 /usr/sbin/httpd
121288 /usr/sbin/httpd
94768 /usr/sbin/httpd
115644 /usr/sbin/httpd
share|improve this question
    
You should provide more information on what you think is "heavy" memory usage - could you update your post with process information during a low-memory condition? –  the-wabbit Oct 8 '11 at 12:46
    
@syneticon-dj: more info added (partial top and RSS info on httpd). It's not at its peak memory usage at the moment but after about a day it starts to rack up points. let me know if there are other relevant info needed. thanks. –  lamp_scaler Oct 8 '11 at 13:25

1 Answer 1

If you see the memory usage increasing over time, it might be a memory leak somewhere. You already tried reducing the "MaxRequestsPerChild" parameter - try tuning it down even more within a controlled environment to see if the total memory stops growing. If it does, you are probably seeing either your PHP code, the PHP interpreter or possibly the PHP cache module leaking memory.

If you see memory leaking, you could try some of the suggestions given for this question on how to debug a memory leak with Apache/PHP.

share|improve this answer
    
I see that error_log() was utilized to see how much RSS was allocated. How does one actually do that with error_log()? Does it show up automatically in strace? –  lamp_scaler Oct 8 '11 at 14:41

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.