Re: [cobalt-users] Memory Usage

[Greg Hewitt-Long <cobaltusers@xxxxxxxxxxxxxxxxxxx>]

>On Thu, 5 Dec 2002, Jalon Q. Zimmerman wrote:
>
>> On Thu, 5 Dec 2002 18:20:43 -0600, Todd W wrote
>> > Is there a way to tell what is taking the most memory as I rebooted
>> > my server and only a day later I am back up to 45% of 512mb.  After
>> > I reboot I am usually around 15%.  This is starting to make me angry
>> > because I think it might be a program of a clients.
>>
>> Todd,
>>
>> I am surprised that its only at 45%. On a Unix/Unix-like OS, it is normal to see apps
>> reserve 90% of the available RAM and free it as others need it.
>>
>> What you would want to watch is when the swap starts being used. Then you are out of
>> RAM.
>>
>> Thats a real simple explanation - I'm sure others could explain it in more detail.
>>
>
>The server pust most memory into 'buffers'
>
>ssh into the server and run free
>look at -/+ buffers/cache
>
>Gerald


I'm trying to get my head round a recent massive increase in the size of our httpd processes - this particularly seems to have affected a RAQ3 which hasn't had any patches or software applied for a few weeks... see the `top` below:

[admin@svr admin]$ top
 11:29am  up 2 days, 19:59,  3 users,  load average: 0.79, 0.80, 5.05
55 processes: 53 sleeping, 2 running, 0 zombie, 0 stopped
CPU states:  9.2% user,  9.9% system,  0.0% nice, 80.7% idle
Mem:   192748K av,  152556K used,   40192K free,  179508K shrd,   11896K buff
Swap:  655816K av,  293056K used,  362760K free                   64988K cached

  PID USER     PRI  NI  SIZE  RSS SHARE STAT  LIB %CPU %MEM   TIME COMMAND
 8198 highlyst  18   0  1648 1648   980 S       0  9.2  0.8   0:00 nq000002.pl
 8155 admin      8   0   880  880   688 R       0  4.3  0.4   0:00 top
 7429 httpd      7   0  278M  13M 11840 R       0  1.3  7.0   0:02 httpd
 7438 httpd      1   0  278M  13M 11880 S       0  1.3  6.9   0:00 httpd
 7403 httpd      1   0  278M  13M 11868 S       0  0.6  7.0   0:00 httpd
 8158 httpd      1   0  278M  13M 12692 S       0  0.6  6.9   0:00 httpd
 6574 admin      0   0  1564 1544  1352 S       0  0.3  0.8   0:00 sshd
 7406 httpd      0   0  278M  13M 12024 S       0  0.3  6.9   0:03 httpd
 7428 httpd      1   0  278M  13M 11876 S       0  0.3  7.0   0:00 httpd
 7448 httpd      0   0  278M  14M 11808 S       0  0.3  7.5   0:01 httpd
    1 root       0   0   120   68    52 S       0  0.0  0.0   0:35 init
    2 root       0   0     0    0     0 SW      0  0.0  0.0   0:15 kflushd
    3 root       0   0     0    0     0 SW      0  0.0  0.0   0:57 kupdate
    4 root       0   0     0    0     0 SW      0  0.0  0.0   0:00 kpiod
    5 root       0   0     0    0     0 SW      0  0.0  0.0   1:03 kswapd
    6 root     -20 -20     0    0     0 SW<     0  0.0  0.0   0:00 mdrecoveryd
 2677 root       0   0   216  160   136 S       0  0.0  0.0   2:03 syslogd

Now... ignore the nq000002.pl process, they only hang around for about 2 seconds - these httpd processes at 278M of usage each are the "problem" - these only seem to have grown to this size in recent weeks - any ideas?

`free` yields this:

[admin@svr admin]$ free
             total       used       free     shared    buffers     cached
Mem:        192748     160376      32372     241664      12148      65636
-/+ buffers/cache:      82592     110156
Swap:       655816     293056     362760


thanks

Greg

ps - the only thing I've done lately is REDUCE the number of objects an httpd process can serve before quitting.

-- 
http://www.webyourbusiness.com/
Providers of E-Commerce Software &
Web Design Consultancy and Services.
PH: (970) 266-0195   FAX: (970) 266-0158