2.4 CPU usage unusually high

[Guest]

[wjstone]

Just installed 2.4 on a brand new Foxconn Atom board with 1 gig of RAM.  With no services turned on except for DHCP the CPU usage on both processors is sitting at 97%, RAM usage is 9%.   Anyone else having this problem?  Not sure why the CPU would be so high.

[wjstone]

Followed the directions here http://efwsupport.com/index.php?topic=1341.msg3230#msg3230 and that seems to have solved it.   Currently i'm using 15% of one cpu and 9% of the other.

[arminf]

WJstone

help me please along the way

use console
Edit the /usr/local/bin/rrdfix.sh  script and remove the /var/ntop from the DIRS variable.
run rrdfix.sh script

thats it or did i miss something?
thanks!

[wjstone]

that should do it,  I didn't notice a difference until I rebooted the firewall. 

Good luck

[arminf]

top - 13:22:23 up 24 min,  1 user,  load average: 6.30, 5.23, 4.06
Tasks: 136 total,   1 running, 135 sleeping,   0 stopped,   0 zombie
Cpu(s):  5.5%us,  1.5%sy,  0.0%ni,  1.0%id, 92.1%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   2067104k total,   478012k used,  1589092k free,    11652k buffers
Swap:   524280k total,        0k used,   524280k free,   112948k cached

what is "wa" which is most of time on +90%?

CPU is still on 100% most of the time.

IPS switched off
InternZone FW traffic Green to Green 60 MB
IPS switched on
InternZone FW traffic Green to Green 3.95 MB

InterZone traffice options "allow", "allow with IPS" -> does not change the result.
ONLY switching IPS off...

[arminf]

Still high CPU

rrdfix didnt worked for me.

WA is taking most of ressources. can somebody explain what it is?

thanks Gents!

login as: root
root@192.168.1.1's password:
root@gateway:~ # top
top - 15:42:03 up 11 min,  1 user,  load average: 3.94, 3.60, 2.02
Tasks: 136 total,   1 running, 135 sleeping,   0 stopped,   0 zombie
Cpu(s):  5.2%us,  1.3%sy,  0.0%ni, 21.6%id, 71.9%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   2067104k total,   450924k used,  1616180k free,     9772k buffers
Swap:   524280k total,        0k used,   524280k free,   110412k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 5175 root      20   0  132m  29m 3248 S   11  1.5   0:53.46 emi
 4702 root      20   0  102m 2504 1248 S    1  0.1   0:09.09 collectd
    6 root      15  -5     0    0    0 S    0  0.0   0:00.09 ksoftirqd/1
 2679 root      20   0  2524 1156  540 S    0  0.1   0:01.06 syslog-ng
 3765 root      20   0 17456 5660  952 S    0  0.3   0:00.53 monit
 6450 root      20   0     0    0    0 D    0  0.0   0:00.11 pdflush
 6872 nobody    20   0  5700 2172 1412 S    0  0.1   0:00.45 httpd
 7386 root      20   0  2000 1060  784 R    0  0.1   0:00.07 top
    1 root      20   0  1508  568  480 S    0  0.0   0:00.68 init
    2 root      15  -5     0    0    0 S    0  0.0   0:00.00 kthreadd
    3 root      RT  -5     0    0    0 S    0  0.0   0:00.01 migration/0
    4 root      15  -5     0    0    0 S    0  0.0   0:28.38 ksoftirqd/0
    5 root      RT  -5     0    0    0 S    0  0.0   0:00.00 migration/1
    7 root      15  -5     0    0    0 S    0  0.0   0:00.85 events/0
    8 root      15  -5     0    0    0 S    0  0.0   0:00.00 events/1
    9 root      15  -5     0    0    0 S    0  0.0   0:00.01 khelper
   61 root      15  -5     0    0    0 S    0  0.0   0:00.40 kblockd/0

[arminf]

1. us -> User CPU time: The time the CPU has spent running users’ processes that are not niced.
2. sy -> System CPU time: The time the CPU has spent running the kernel and its processes.
3. ni -> Nice CPU time: The time the CPU has spent running users’ proccess that have been niced.
4. wa -> iowait: Amount of time the CPU has been waiting for I/O to complete.
5. hi -> Hardware IRQ: The amount of time the CPU has been servicing hardware interrupts.
6. si -> Software Interrupts.: The amount of time the CPU has been servicing software interrupts.

can i somewhere check which I/os are waiting?

Some say its a disk issue..

[wjstone]

What sort of disk are you using?  Some kind of solid state, or compact flash?

[arminf]

Hi WJStone

its a cf card 133 DMA speed.

was not a problem with 2.3
But i am thinking to put a SSD in if this is  the case.

Its mostly the %wa state. Funny is as soon as i start copying from green to green wa% goes down.
thanks!

[arminf]

As soon the box is rebooted and got "nothing to do" WA state goes high.

When i start copy, streaming, downloading the state goes down. as soon as i finished it raises up again

here CPU states

right after reboot
Cpu(s): 89.4%us, 10.1%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.2%hi,  0.3%si,  0.0%st
Mem:   2067104k total,   325384k used,  1741720k free,     8112k buffers
Swap:   524280k total,        0k used,   524280k free,   138320k cached

6 Minutes uptime WA starts
top - 01:06:32 up 6 min,  1 user,  load average: 3.14, 2.77, 1.32
Tasks: 136 total,   1 running, 135 sleeping,   0 stopped,   0 zombie
Cpu(s):  4.8%us,  1.8%sy,  0.0%ni,  3.8%id, 89.6%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   2067104k total,   457144k used,  1609960k free,     9852k buffers
Swap:   524280k total,        0k used,   524280k free,   117272k cached


while copy green to green (SNORT enabled = 6 MB speed :-( )
top - 17:36:52 up 10 min,  1 user,  load average: 6.38, 4.11, 2.17
Tasks: 152 total,   3 running, 147 sleeping,   0 stopped,   2 zombie
Cpu(s): 85.2%us,  7.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.5%hi,  7.3%si,  0.0%st
Mem:   2067104k total,   476232k used,  1590872k free,    10676k buffers
Swap:   524280k total,        0k used,   524280k free,   121944k cached


right after copy
top - 17:40:48 up 14 min,  1 user,  load average: 2.95, 4.62, 2.91
Tasks: 130 total,   1 running, 129 sleeping,   0 stopped,   0 zombie
Cpu(s): 50.0%us,  1.0%sy,  0.0%ni, 37.7%id, 11.3%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   2067104k total,   549624k used,  1517480k free,    11872k buffers
Swap:   524280k total,        0k used,   524280k free,   125580k cached

changing SNORT rule (deactivate TOR ruleset)
top - 17:44:41 up 18 min,  1 user,  load average: 1.72, 2.92, 2.58
Tasks: 133 total,   5 running, 128 sleeping,   0 stopped,   0 zombie
Cpu(s): 96.2%us,  3.8%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   2067104k total,   455636k used,  1611468k free,    12568k buffers
Swap:   524280k total,        0k used,   524280k free,   126116k cached


After SNORT rule change and reboot while copy (42 MB speed :-) )
top - 18:00:52 up 10 min,  1 user,  load average: 1.48, 1.69, 1.11
Tasks: 131 total,   2 running, 129 sleeping,   0 stopped,   0 zombie
Cpu(s): 48.1%us,  1.8%sy,  0.0%ni,  0.0%id,  0.0%wa,  1.5%hi, 48.6%si,  0.0%st
Mem:   2067104k total,   437776k used,  1629328k free,     9816k buffers
Swap:   524280k total,        0k used,   524280k free,   100832k cached

[arminf]

top - 18:39:59 up 49 min,  1 user,  load average: 5.47, 4.88, 3.79
Tasks: 135 total,   1 running, 134 sleeping,   0 stopped,   0 zombie
Cpu(s):  6.3%us,  1.8%sy,  0.0%ni,  0.0%id, 91.9%wa,  0.0%hi,  0.0%si,  0.0%st
Mem:   2067104k total,   473128k used,  1593976k free,    16108k buffers
Swap:   524280k total,        0k used,   524280k free,   119848k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
5091 root      20   0  132m  30m 3248 S   14  1.5   6:23.97 emi
 4709 root      20   0  102m 2672 1248 S    2  0.1  25:34.11 collectd

wwhat is emi? this could be the one makng trouble...

[arminf]

Tasks: 134 total,   1 running, 132 sleeping,   0 stopped,   1 zombie
Cpu(s):  6.0%us,  2.3%sy,  0.0%ni, 16.1%id, 75.5%wa,  0.0%hi,  0.2%si,  0.0%st
Mem:   2067104k total,   490964k used,  1576140k free,    28592k buffers
Swap:   524280k total,        0k used,   524280k free,   123384k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
5091 root      20   0  132m  30m 3260 S   13  1.5   8:46.03 emi
 4709 root      20   0  103m 2704 1248 S    3  0.1  26:38.66 collectd

still same both processes are open.

somebody has a clue?


thanks!!


still same....

top - 20:56:52 up  3:06,  1 user,  load average: 7.11, 5.25, 4.48
Tasks: 138 total,   1 running, 135 sleeping,   0 stopped,   2 zombie
Cpu(s):  6.1%us,  2.2%sy,  0.0%ni,  0.0%id, 91.6%wa,  0.0%hi,  0.2%si,  0.0%st
Mem:   2067104k total,   578684k used,  1488420k free,    31704k buffers
Swap:   524280k total,        0k used,   524280k free,   124684k cached

  PID USER      PR  NI  VIRT  RES  SHR S %CPU %MEM    TIME+  COMMAND
 4709 root      20   0  103m 2752 1248 S    3  0.1  27:11.50 collectd
 5091 root      20   0  132m  30m 3260 S   14  1.5  11:50.74 emi

[arminf]

ok

EMI = endian management interface (webserver GUI) ??

collectd
http://bugs.endian.com/view.php?id=2585

collectd is a lighweight, modular daemon that collects system performance data
Dashboard?

so better close the admin gui to safe resources ?!? :-S

collectd is running most of time (schedule for perfomance checking?) together with monit

[arminf]

High Percentage in %id CPU State

Programs which make use of CPU Idle Time mean that they run at a low priority so as not to impact programs that run at normal priority. Many programs that use CPU idle time cause the CPU to always be 100% utilized, so that the time spent where the CPU would have been idle is instead spent performing useful computations. This generally causes the CPU to consume more power as most modern CPUs can enter power-save modes when they are idle.

Not good for an ATOM CPU....

[arminf]

top - 07:58:47 up 14:08,  1 user,  load average: 6.44, 5.58, 3.66
Tasks: 136 total,   1 running, 135 sleeping,   0 stopped,   0 zombie
Cpu(s):  6.1%us,  2.5%sy,  0.0%ni,  0.0%id, 91.2%wa, 0.0%hi,  0.2%si,  0.0%st
Mem:   2067104k total,   657752k used,  1409352k free,    99532k buffers
Swap:   524280k total,        0k used,   524280k free,   187856k cached


still high cpu consumption..

any idea?