Memory Ram too high in HTTPD processes

Question

I management two dedicated servers. I use Centos 6 with Plesk Panel. Update: I use Apache with mod_php.

On the first server I have a site in Wordpress.

I have noticed that my httpd processes take up more memory each time. Besides the 'top' command, I use the following command to find out:

ps -ylC httpd --sort:rss

Image:

Pay attention to the RSS column. The httpd processes from 13MB to 127MB occupy in ram.

Installed apache modules are:

# Httpd l 
Compiled in modules: 
   core.c 
   prefork.c 
   http_core.c 
   mod_so.c

On the second server I have a website PHPBB.

In this case the involved processes like httpd. All occupy about 85MB.

Image:

I've read that should occupy 20MB. How I can profiling or optimize this? With what tool? Xhprof I tried, but it says that the memory used is less than that then use really ...

My memory are intensive in my servers. This is a big problem.

Update:

Server 1 Meminfo output:

#cat /proc/meminfo
MemTotal:        5969120 kB
MemFree:          625720 kB
Buffers:          218424 kB
Cached:          3289320 kB
SwapCached:        61020 kB
Active:          2832624 kB
Inactive:        2066428 kB
Active(anon):    1111532 kB
Inactive(anon):   343184 kB
Active(file):    1721092 kB
Inactive(file):  1723244 kB
Unevictable:           0 kB
Mlocked:               0 kB
SwapTotal:       1959920 kB
SwapFree:        1789764 kB
Dirty:              1044 kB
Writeback:             0 kB
AnonPages:       1348048 kB
Mapped:            53712 kB
Shmem:             63408 kB
Slab:             379972 kB
SReclaimable:     220388 kB
SUnreclaim:       159584 kB
KernelStack:        2064 kB
PageTables:        28848 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:     4944480 kB
Committed_AS:    3148736 kB
VmallocTotal:   34359738367 kB
VmallocUsed:       24764 kB
VmallocChunk:   34359713428 kB
HardwareCorrupted:     0 kB
AnonHugePages:         0 kB
HugePages_Total:       0
HugePages_Free:        0
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB
DirectMap4k:     6291456 kB
DirectMap2M:           0 kB

Server 1 pmap of the httpd most memory processes (127mb and 118mb):

# pmap 11737
11737:   /usr/sbin/httpd
 total                0K
# pmap 12099
12099:   /usr/sbin/httpd
 total                0K

Server 2 Meminfo output:

# cat /proc/meminfo
MemTotal:       10088968 kB
MemFree:         1773488 kB
Buffers:          237064 kB
Cached:          5752804 kB
SwapCached:        38764 kB
Active:          4516332 kB
Inactive:        2738344 kB
Active(anon):    1028388 kB
Inactive(anon):   267156 kB
Active(file):    3487944 kB
Inactive(file):  2471188 kB
Unevictable:           0 kB
Mlocked:               0 kB
SwapTotal:       1959920 kB
SwapFree:        1883656 kB
Dirty:              4444 kB
Writeback:             0 kB
AnonPages:       1226476 kB
Mapped:            37020 kB
Shmem:             30736 kB
Slab:             862964 kB
SReclaimable:     762056 kB
SUnreclaim:       100908 kB
KernelStack:        2424 kB
PageTables:        57564 kB
NFS_Unstable:          0 kB
Bounce:                0 kB
WritebackTmp:          0 kB
CommitLimit:     7004404 kB
Committed_AS:    2753212 kB
VmallocTotal:   34359738367 kB
VmallocUsed:       42048 kB
VmallocChunk:   34359642628 kB
HardwareCorrupted:     0 kB
AnonHugePages:         0 kB
HugePages_Total:       0
HugePages_Free:        0
HugePages_Rsvd:        0
HugePages_Surp:        0
Hugepagesize:       2048 kB
DirectMap4k:    10485760 kB
DirectMap2M:           0 kB

Server 2 httpd pmap:

# pmap 29772
29772:   /usr/sbin/httpd
 total                0K

Hi Cookboy. You are asking quite a bit in this question, particularly in the paragraph that begins with "But this is not normal right?", and I imagine that is why it was downvoted. Consider **[edit]ing** your question to focus it on a single issue that can ideally be answered authoritatively within a few paragraphs. Also consider that the scope of this site is basically system administration in a professional context; if you are currently running a professional system where 100 MB of RAM matters that much, you should at least elaborate on why that is so, because to most that's margin of error. — user, Aug 18 '14 at 11:39
Thank you so much. I will change my question, sorry for this, is my first question in ServerFault :) Thanks! — Cookboy, Aug 18 '14 at 11:41
We all start out somewhere! Keep in mind the scope of each site (see the respective help center, particularly [What topics can I ask about here?](http://serverfault.com/help/on-topic) and [What types of questions should I avoid asking?](http://serverfault.com/help/dont-ask)) and that the Q&A format works best when a single, authoritative answer is possible, and you'll likely do fine. If you are unsure, ask on the respective site's Meta site first (for example [meta]) whether a question on a given topic would be on topic on the site or not (but don't ask the actual question there). — user, Aug 18 '14 at 11:43
OK, I understand. I edited my question. It is right now? Thanks again! — Cookboy, Aug 18 '14 at 11:47
Are you using `mod_php` or fpm? That matters since if you're using the Apache module memory usage like this is normal for the PHP engine. — Nathan C, Aug 18 '14 at 12:09
Hello, phpinfo() tells me that I have loaded the module mod_php5. If I'm not mistaken this means I'm using mod_php, right? So this is normal behavior? Thanks you for your help. — Cookboy, Aug 18 '14 at 12:29
I think you are not asking the right question. How did you determined that you have a memory usage problem? Could you add the output of `cat /proc/meminfo`? You can see how the virtual memory is mapped with `pmap $PID` or `cat /proc/$PID/maps`. Where $PID is the process ID that is using unexpected amounts of memory. — Mircea Vutcovici, Aug 18 '14 at 15:13
I edited the first post with the information you ask. I think I have a memory problem because my httpd processes use much more RAM than I think it should be used. I think they should use around 20mb. But in my case using 85MB or 120mb. Let me know if you can optimize memory usage of httpd processes and how. I had to buy memory and spend money, I prefer to optimize these processes if possible. Thank you very much. — Cookboy, Aug 19 '14 at 12:40

score 5 · Answer 1 · answered Aug 18 '14 at 13:31

5

You're using mod_php, so Apache is also handling your PHP workloads at the same time. Since PHP needs lots of RAM (well, most like this do, including mod_perl), you'll see higher RAM usage than serving just static files.

answered Aug 18 '14 at 13:31

Nathan C

14,901
4
42
62

Yes, I understand this. So this is not a problem? Is it normal to high memory usage of my httpd processes? So I just have to buy more RAM if I need it? Thanks! – Cookboy Aug 19 '14 at 12:43
@Cookboy Yes, more RAM. You can also try switching to `php-fpm` instead, but this is beyond the scope of your question. With it though, you can finely tune FPM to use less memory at the expense of performance if you encounter high load. Similar can be done under `mod_php` by limited the number of children HTTPD can spawn. – Nathan C Aug 19 '14 at 17:30
OK, but why do some processes use 13mb or 80mb and other 127MB? I do not understand. I need to know this information anyway. Is this related with the php client calls? How? Thanks again. – Cookboy Aug 21 '14 at 10:58
@Cookboy It really depends. I'm not very familiar with the internal workings of `mod_php`, but I'm assuming it's based on how much processing it does and how much is cached data in PHP. Usually the parent Apache process recycles high-memory children on a regular basis. – Nathan C Aug 21 '14 at 18:24

Memory Ram too high in HTTPD processes

1 Answers1