Does my Centos server leak memory?

Question

I'm getting some conflicting information regarding the memory use of a virtualized server. The most alarming information is this:

Note that the server has only 4GB of RAM allocated, even though committed shows 12GB.

Also, if i run ps aux | grep httpd I get the following:

root      1566  0.1  9.5 580392 375540 ?       Ss   Jun18  96:15 /usr/local/apache/bin/httpd -k start -DSSL
root      4212  0.0  9.4 580388 371948 ?       S    08:01   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody    7674  0.6  9.5 580392 373600 ?       S    08:26   0:02 /usr/local/apache/bin/httpd -k start -DSSL
nobody   10894  0.6  9.5 580532 373508 ?       S    08:28   0:01 /usr/local/apache/bin/httpd -k start -DSSL
nobody   11668  0.6  9.5 580392 373600 ?       S    08:29   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   11669  0.6  9.5 580392 373512 ?       S    08:29   0:01 /usr/local/apache/bin/httpd -k start -DSSL
nobody   11975  0.6  9.5 580392 373492 ?       S    08:29   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   12108  0.7  9.5 580392 373492 ?       S    08:29   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   12993  0.6  9.5 580392 373492 ?       S    08:30   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   13080  0.6  9.5 580392 373492 ?       S    08:30   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   13211  0.6  9.5 580392 373492 ?       S    08:30   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   13496  0.6  9.5 580392 373520 ?       S    08:30   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   13497  0.6  9.5 580392 373492 ?       S    08:30   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   13829  0.6  9.5 580392 373492 ?       S    08:30   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   13858  0.6  9.5 580392 373492 ?       S    08:30   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   13963  0.6  9.5 580392 373484 ?       S    08:30   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   14070  0.7  9.5 580392 373492 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   14290  0.6  9.5 580392 373492 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   14404  0.6  9.5 580392 373484 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   14745  0.6  9.5 580392 373492 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   14850  0.6  9.5 580392 373484 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   14957  0.5  9.5 580392 373484 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15061  0.6  9.5 580392 373492 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15174  0.8  9.5 580392 373492 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15228  0.6  9.5 580392 373484 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15263  0.4  9.5 580392 373484 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15264  0.5  9.5 580392 373484 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15290  0.7  9.5 580392 373484 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15291  0.5  9.5 580392 373484 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15294  0.4  9.5 580392 373484 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15296  0.6  9.5 580392 373492 ?       S    08:31   0:00 /usr/local/apache/bin/httpd -k start -DSSL
nobody   15401  0.5  9.5 580392 373484 ?       S    08:32   0:00 /usr/local/apache/bin/httpd -k start -DSSL
root     15506  0.0  0.0 103228   840 pts/0    S+   08:32   0:00 grep httpd

This totals up to over 294.5% of memory used.

top gives similar results too.

However, if I run free -m I get this:

             total       used       free     shared    buffers     cached
Mem:          3831       3647        184          0        257       2358
-/+ buffers/cache:       1031       2799
Swap:         4031         29       4002

This seems OK to me, no sign of overcommitted memory.

The server is relatively quick, there's no unusual downtime/sluggishness, and I'd like to keep it that way. Are the abnormal readings just an anomaly, or something more serious?

Answer 1

You have enough memory. You would have problem with memory if your used memory rizes to max amount and buffers and cached will be very low, using of SWAP will rize then you probably will have problems with free RAM. From your post it's everything ok....

This is simple explanation what commited memory is:

Committed_AS: An estimate of how much RAM you would need to make a 99.99% guarantee that there never is OOM (out of memory) for this workload. Normally the kernel will overcommit memory. That means, say you do a 1GB malloc, nothing happens, really. Only when you start USING that malloc memory you will get real memory on demand, and just as much as you use. So you sort of take a mortgage and hope the bank doesn't go bust. Other cases might include when you mmap a file that's shared only when you write to it and you get a private copy of that data. While it normally is shared between processes. The Committed_AS is a guesstimate of how much RAM/swap you would need worst-case.

Here is hole explanation http://www.redhat.com/advice/tips/meminfo.html

So you don't have problems with memory everythings is ok those readings are normal.

Answer 2

There is a shared memory usage with your apache model, this is fine. ps/top won't show you exact usage, that's not an easy task. Try http://www.pixelbeat.org/scripts/ps_mem.py