SatSig topic:

x3 cpu usage

(Read 6621 times)

Sep 30^th, 2008 at 2:56pm

my x3 cpu usage always rises.
i know from my ex-3100 it should always reside in the 1-10% zone.

when should i start to worry?

Oct 1^st, 2008 at 1:59am

You are one of few using the new iDirect S2 modem. I dont have anything to contribute as I lack experience with that modem, but I am certainly following along with your posts.

Oct 1^st, 2008 at 8:36am

this is what its showing now with hardly any traffic going in or out.

Quote:

[RMT:6666] admin@telnet:::ffff:127.0.0.1;1024
> cpu
CPU Load : 7%

[RMT:6666] admin@telnet:::ffff:127.0.0.1;1024
> ps
PID Uid VmSize Stat Command
1 root 508 S init [2]
2 root SW [keventd]
3 root SWN [ksoftirqd_CPU0]
4 root SW [kswapd]
5 root SW [bdflush]
6 root SW [kupdated]
8 root SW [mtdblockd]
42 root SWN [jffs2_gcd_mtd2]
61 root DW [ixp425_csr]
63 root SW [ixp425 ixp0]
65 root SW [ixp425 ixp1]
216 root 564 S /sbin/syslogd -m 0
219 root 544 S /sbin/klogd
240 root 1408 S /usr/sbin/sshd
365 root 16832 S /usr/sbin/falcon
371 root 16832 S /usr/sbin/falcon
372 root 16832 S /usr/sbin/falcon
373 root 936 S /usr/sbin/falcon_monitor
381 root 308 S /bin/ptylogin
382 root 720 S -ash
383 root 304 S /bin/ptylogin
390 root 632 S telnet 0
391 root 648 R /bin/ps

[RMT:6666] admin@telnet:::ffff:127.0.0.1;1024

start doing a bit of browsing and you'll see cpu report 30% usage almost immediately!

I think it has to do with the fact that i am using a 240 subnet because i had a similar incident with a 3100 in the past and we resolved it by assigning a public ip ...

on a side note if run the "top" command logged in as root through console i can see that falcon consumes 16M of memory and near almost all the cpu.

iDirect really should take part in optimizing the falcon code.

Quote:

Mem: 36340K used, 26304K free, 0K shrd, 0K buff, 12048K cached
Load average: 1.18, 1.06, 1.01 (State: S=sleeping R=running, W=waiting)

PID USER STATUS RSS PPID %CPU %MEM COMMAND
365 root R 16M 1 8.7 26.8 falcon
392 root R 712 382 1.9 1.1 busybox
371 root S 16M 365 0.0 26.8 falcon
372 root S 16M 371 0.0 26.8 falcon
240 root S 1408 1 0.0 2.2 sshd
373 root S 936 1 0.0 1.4 falcon_monitor
382 root S 720 381 0.0 1.1 ash
216 root S 564 1 0.0 0.9 syslogd
219 root S 544 1 0.0 0.8 klogd
1 root S 508 0 0.0 0.8 init
381 root S 308 1 0.0 0.4 ptylogin
383 root S 304 381 0.0 0.4 ptylogin
6 root SW 0 1 0.0 0.0 kupdated
2 root SW 0 1 0.0 0.0 keventd
3 root SWN 0 1 0.0 0.0 ksoftirqd_CPU0
4 root SW 0 1 0.0 0.0 kswapd
5 root SW 0 1 0.0 0.0 bdflush
8 root SW 0 1 0.0 0.0 mtdblockd
42 root SWN 0 1 0.0 0.0 jffs2_gcd_mtd2

i am also dying to find out why falcon is running in three instances. (365, 371, 372) each consuming 16M of memory (totaling 48M).
The whole modem only has 64 so they're nearly chocking it!!!
then you get a piece of junk xscale arm processory running at ONLY 266Mhz!
wow! talk about cheap!!!!!

I wonder if i can overclock this thing.... hmm...

Processor : XScale-IXP4xx/IXC11xx rev 2 (v5b)
BogoMIPS    : 266.24
Features    : swp half thumb fastmult edsp

Hardware    : iDirect Phoenix Platform
Revision    : 0000
Serial    : 0000000000000000

Oct 1^st, 2008 at 10:14am

Currently i am going to configure a large batch. Therefore i shall post my findings. Have you kill the extra X 2 PID falcon. If so what happened ?

Have you loaded iperf ? Do you get the same type of mem usage ? Is it http traffic only ?

Oct 1^st, 2008 at 11:00am

no i have not killed the extra 2.

i will kill them later today and will update my findings.

by the way this is 9.0.0.5

i have not used iperf yet. will also do that later.

Oct 1^st, 2008 at 7:54pm

ok i killed the other two

now top only shows 1 running and the system seems to be ok.

Mem: 36532K used, 26112K free, 0K shrd, 0K buff, 12048K cached
Load average: 1.02, 1.02, 1.00 (State: S=sleeping R=running, W=waiting)

PID USER    STATUS RSS PPID %CPU %MEM COMMAND
365 root    R    16M    1 11.1 26.9 falcon
401 root    R    760 382 0.5 1.2 busybox
240 root    S 1408    1 0.0 2.2 sshd
373 root    S    936    1 0.0 1.4 falcon_monitor
382 root    S    824 381 0.0 1.3 ash
216 root    S    564    1 0.0 0.9 syslogd
219 root    S    544    1 0.0 0.8 klogd
1 root    S    508    0 0.0 0.8 init
381 root    S    308    1 0.0 0.4 ptylogin
383 root    S    304 381 0.0 0.4 ptylogin
6 root    SW    0    1 0.0 0.0 kupdated
2 root    SW    0    1 0.0 0.0 keventd
3 root    SWN    0    1 0.0 0.0 ksoftirqd_CPU0
4 root    SW    0    1 0.0 0.0 kswapd
5 root    SW    0    1 0.0 0.0 bdflush
8 root    SW    0    1 0.0 0.0 mtdblockd
42 root    SWN    0    1 0.0 0.0 jffs2_gcd_mtd2
61 root    DW    0    1 0.0 0.0 ixp425_csr
63 root    SW    0    1 0.0 0.0 ixp425 ixp0

i am not too familiar on using iperf. can you advise on how to set it up for this purpose?
thx

Oct 2^nd, 2008 at 3:44am

I have never used iperf either. I prefer IXChariot.

Still interested in any feedback from anyone who knows how to set it up and is fluent in using it. (iperf)

Oct 2^nd, 2008 at 2:56pm

well in order for iperf to work you'd have to have an iperf on the terminal but there isnt any.

i searched the modem bin directories and found no iperf.

whoever did it either had a iperf binary for the unit or had an image that came with iperf.

Oct 2^nd, 2008 at 5:12pm

If you have a support contract with iDirect, they may be able to supply you with a version of iperf called iperf.infiniti

it is recommended that this is not used in standard set up as it increases the CPU load. So they may ask you why you want it. I believe it is really for the guys there to test and run on courses etc.

Oct 2^nd, 2008 at 5:16pm

and it will run on the X3 also BTW

Oct 2^nd, 2008 at 8:49pm

ya i know it'll run on x3.

the ONLY difference X3 has is the dvbs2 component. oh and the board is now smaller whereas in the 3100 the whole case filled the board. now the x3 has a smaller board but the casing is the same size as the 3100.

i am looking for an arm compiled iperf binary. i may end up compiling one myself.

Oct 4^th, 2008 at 9:22am

cylent,
The question you had about the CPU utilization during browsing and using the top command is normal, as I can see from your screen shot the CPU usage is low.
Secondly the extra 2 falcon process’s are threads of the parent (first)falcon process, and not using 16M each, all 3 are consuming 16M, this can be verified by using the free command.

Good Luck.