CPU usage increases suddenly after a few minutes

[Tyrann]

Hi,

i have a big problem with my Counter Strike Source Dedicated Server (no addons, just vanilla), the first few minutes the server runs at a CPU usage of ~ 35 % at 12 players but then the cpu usage increases rapidly to 97 % so it lags. I tried differend ways to solve this problem:

1. i looked at my server config (i use your standard config of the forums) nothing happend.
2. i looked at my logs of the server (-debug is on) everything seems to be ok.
3. the srcds_run autodetects my intel celeron 2.0 ghz and uses the i686 library so i tried to change it (-binary ./srcds_i486), but also with no effect on my problem.
4. i reinstalled steam and everything is upgraded -> nothing happend


i hope you can help me with my problem because i am helpless.

Code:

Some Stystem Information:
processor       : 0
vendor_id       : GenuineIntel
cpu family      : 15
model           : 2
model name      : Intel(R) Celeron(R) CPU 2.00GHz
stepping        : 7
cpu MHz         : 2000.550
cache size      : 128 KB
fdiv_bug        : no
hlt_bug         : no
f00f_bug        : no
coma_bug        : no
fpu             : yes
fpu_exception   : yes
cpuid level     : 2
wp              : yes
flags           : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe cid
bogomips        : 3948.54

512 mb ddr ram

Kernel 2.6.3

Debian Testing (newest glibc for debian (GNU libc) 2.3.2)

if you need more informations about the server i will post it :)

[theansaname]

can you screenshot your tops command b4 and after the spike?

[Tyrann]

Code:

top - 19:36:11 up 19:52,  1 user,  load average: 0.67, 0.68, 0.46
Tasks:  49 total,   2 running,  47 sleeping,   0 stopped,   0 zombie
Cpu(s): 92.4% us,  3.3% sy,  0.0% ni,  4.3% id,  0.0% wa,  0.0% hi,  0.0% si
Mem:    512396k total,   324488k used,   187908k free,    83164k buffers
Swap:  1967952k total,        0k used,  1967952k free,    89628k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
3058 tfed      18   0  129m  59m 8188 R 94.9 12.0  19:44.04 srcds_i686
12367 root      16   0  2060 1052  844 R  0.3  0.2   0:01.25 top
    1 root      16   0  1500  516  456 S  0.0  0.1   0:00.17 init
    2 root      34  19     0    0    0 S  0.0  0.0   0:00.07 ksoftirqd/0
    3 root       5 -10     0    0    0 S  0.0  0.0   0:00.00 events/0
    4 root       5 -10     0    0    0 S  0.0  0.0   0:00.02 khelper
   16 root      15 -10     0    0    0 S  0.0  0.0   0:00.00 kacpid
   90 root       5 -10     0    0    0 S  0.0  0.0   0:00.08 kblockd/0
  158 root      15   0     0    0    0 S  0.0  0.0   0:00.00 pdflush
  159 root      15   0     0    0    0 S  0.0  0.0   0:00.18 pdflush
  161 root      11 -10     0    0    0 S  0.0  0.0   0:00.00 aio/0
  160 root      17   0     0    0    0 S  0.0  0.0   0:00.13 kswapd0
  749 root      25   0     0    0    0 S  0.0  0.0   0:00.00 kseriod
  798 root      15   0     0    0    0 S  0.0  0.0   0:00.99 kjournald
1048 root      16   0  1556  648  528 S  0.0  0.1   0:00.24 syslogd
1051 root      18   0  2352 1420  444 S  0.0  0.3   0:00.09 klogd
1059 root      16   0  1536  528  464 S  0.0  0.1   0:00.00 inetd
1071 root      25   0  2500 1224 1012 S  0.0  0.2   0:00.00 mysqld_safe
1111 mysql     16   0 64368 6456 2520 S  0.0  1.3   0:00.13 mysqld
1113 root      17   0  1484  496  436 S  0.0  0.1   0:00.00 logger
1143 nobody    16   0  1708  660  576 S  0.0  0.1   0:00.00 oidentd
1148 scanlogd  15   0  1688  648  552 S  0.0  0.1   0:00.54 scanlogd
1160 root      16   0  3464 1536 1300 S  0.0  0.3   0:00.00 sshd
1165 root      16   0  2228 2220 1884 S  0.0  0.4   0:00.08 ntpd
1171 daemon    16   0  1680  636  556 S  0.0  0.1   0:00.00 atd
1174 root      16   0  1760  816  680 S  0.0  0.2   0:00.04 cron
1179 root      17   0  1496  480  420 S  0.0  0.1   0:00.00 getty
1181 root      16   0  1496  480  420 S  0.0  0.1   0:00.00 getty
1182 root      17   0  1496  480  420 S  0.0  0.1   0:00.00 getty
1183 root      16   0  1496  480  420 S  0.0  0.1   0:00.00 getty
1184 root      16   0  1496  480  420 S  0.0  0.1   0:00.00 getty
1186 root      17   0  1496  480  420 S  0.0  0.1   0:00.00 getty
1212 root      16   0  4984 2116 1580 S  0.0  0.4   0:00.03 httpd
1213 nobody    16   0  5816 3396 2084 S  0.0  0.7   0:00.04 httpd
1214 nobody    16   0  5816 3444 2084 S  0.0  0.7   0:00.07 httpd
1215 nobody    16   0  4992 2208 1640 S  0.0  0.4   0:00.00 httpd
1216 nobody    16   0  5812 3392 2084 S  0.0  0.7   0:00.04 httpd
1217 nobody    16   0  5860 3576 2120 S  0.0  0.7   0:00.06 httpd
1218 nobody    16   0  5860 3564 2112 S  0.0  0.7   0:00.05 httpd
1221 tfed      16   0  2664 1208  896 S  0.0  0.2   0:00.02 screen
1222 tfed      15   0  2404 1184  900 S  0.0  0.2   0:00.01 srcds_run
1637 tyrann    16   0  4020 2768 1644 S  0.0  0.5   0:01.41 eggdrop
1850 tyrann    16   0  3484 2216 1580 S  0.0  0.4   0:00.92 eggdrop
2959 nobody    16   0  4992 2204 1640 S  0.0  0.4   0:00.00 httpd
8242 nobody    16   0  4984 2196 1636 S  0.0  0.4   0:00.00 httpd
8243 nobody    16   0  4992 2204 1640 S  0.0  0.4   0:00.00 httpd
8244 nobody    16   0  4992 2196 1632 S  0.0  0.4   0:00.00 httpd
12362 root      16   0 14616 2096 1720 S  0.0  0.4   0:00.17 sshd
12365 root      15   0  2592 1432 1120 S  0.0  0.3   0:00.02 bash

view moments later

[Tyrann]

i have still the same problems

now i use the kernel 2.6.10

i tried my luck with cs 1.6 and hl1/2 deathmatch servers but always the same problem after a few minutes and over 7 - 8 players the cpu usage increases to 98 %. so ... is it a bug from the valve engines with my celeron ? because everything runs smooth before 2 months ...

i still hope someone can help me

i don't want to reinstall everything :/

[theansaname]

do you have an optimization settings?

[Tyrann]

hmm thx i think this is the problem,

how can i add SSE support to my 2.6.10 Kernel ?

i can only choose the CPU nothing else :/

[joltservers]

Same thing going on with me - I have the same kernel and a p4 3.06, and it's happening on other boxes of mine with a variety of CPUs.. Seems like it just started happening very recently - like within the last week or 2.

Let me know if you've had any luck resolving this.

[Tyrann]

i try to reinstall my hole debian server :/ to woody with glib 2.3.2

i will tell you if anything will happen

are u a admin of the jolt co uk servers ?

[joltservers]

cool, keep us posted! I'm wondering if it has to do with either the kernel or a recent update from steam. Thing is it seemed to just start happeneing since around the end of Jan or beginning of Feb.

I do have it running with a variety of kernels and it doesn't seem I'm having the same result everywhere.

I'm going to tar up an install that is particularly bad from one of my servers and put it on another box with an earlier kernel and see what happens. I'll let you know.

Also, are you running manimod?

and no - i'm an admin of joltservers dot com

[Tyrann]

yesterday i was at the data centre and i was shocked that my cpu fan was damaged or something like that, so i think my problem is that my cpu clock turns down after a certain temperature. So i repaired the cpu fan and tryed it again with 16 bots but the problem seems to be still alive because the cpu usage changes wild from 30 % to 97 % back to 50 % and so on ...

now i will try to check if my cpu is damaged or the cpu is still getting to high temps :/

are u also using debian ?

i will post more infos about my system

Code:

tyrann:~# /lib/libc.so.6
GNU C Library stable release version 2.3.2, by Roland McGrath et al.
Copyright (C) 2003 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.
There is NO warranty; not even for MERCHANTABILITY or FITNESS FOR A
PARTICULAR PURPOSE.
Compiled by GNU CC version 3.3.5 (Debian 1:3.3.5-5).
Compiled on a Linux 2.6.0-test7 system on 2005-01-05.
Available extensions:
        GNU libio by Per Bothner
        crypt add-on version 2.1 by Michael Glad and others
        linuxthreads-0.10 by Xavier Leroy
        BIND-8.2.3-T5B
        libthread_db work sponsored by Alpha Processor Inc
        NIS(YP)/NIS+ NSS modules 0.19 by Thorsten Kukuk
Thread-local storage support included.
Report bugs using the `glibcbug' script to <bugs@gnu.org>.

this bug/problem is making me crazy :/

[joltservers]

I'm using fedora core 3 - I'm going to tar and move one of the installs i'm having trouble with over to an FC2 box and see how it behaves.

[Delphis]

Hi,

Same problem here, after 30 mins the cpu usage goes bananas between 25-80 % depending on the amount of players active (not death).

1 CPU P4 3,06
2 Gb mem
Fedora Core 3

1 thing it's running in SMP Multi processor mode due to the hyperthreating cpu, could this be the thing? Because in top i see 25% cpu usage and by the ./srcds it's double the amount.

1 more thing is the cashed memory, it's eating all the memory till there is left only 8mb, why cache? what is the point? maps ok but cache what players do etc. ?


Im rebooting the server during the night to free the memory. don't know what else to do?

Overview of TOP

Code:

top - 14:43:03 up  7:43,  1 user,  load average: 0.40, 0.40, 0.37
Tasks:  59 total,   1 running,  58 sleeping,   0 stopped,   0 zombie
Cpu(s):  6.5% us,  1.0% sy,  0.0% ni, 92.2% id,  0.0% wa,  0.0% hi,  0.3% si
Mem:   2042668k total,  1555900k used,   486768k free,    58944k buffers
Swap:  2088432k total,        0k used,  2088432k free,  1201016k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
4191 observer  15   0  124m  56m 8288 S  9.3  2.8   7:30.20 srcds_i686
3645 observer  15   0  145m  76m 8284 S  2.7  3.8  31:29.40 srcds_i686
3252 root      15   0  2312  588  496 S  1.0  0.0   3:10.87 syslogd
3256 root      15   0  2164  480  416 S  1.0  0.0   5:10.08 klogd
4667 observer  16   0  2976  932  756 R  0.7  0.0   0:00.07 top

Grtz Delphis

[joltservers]

I'm running that exact same setup - same cpu, same OS. I am starting to think it's the 2.6.9/2.6.10 kernels - as my FC2 and RH9 boxes seem ok.

I turned SMP/HT off on my box as well and had the same issue.

I think you're ok on the cached memory though, it's just kind of the way linux deals w/ RAM.

[Delphis]

Well the CPU usage is increasing when the "free" memory becomes less and less.

Ive updated FD from core2 to core3 due to this problems but there is no difference.

[joltservers]

any mods installed? mani mod? on one of my boxes, it seems to be running fine running this kernel: 2.6.5-1.358smp