Re: Traffic counts wrongly in OUTPUT.

Re: Traffic counts wrongly in OUTPUT.

[Daniel]

On Sat, 5 Mar 2005 17:46:03 +1030, <deniska@westnet.com.au> wrote:

> Hi,
>
> My counter in OUTPUT grows with >10MBytes/sec but I have only 64kbit of
> bandwidth. I have QoS shaping outgoing traffic, so, my guess is, the
> kernel sends packets as quick as possible but most of them are
> dropped in a queue by QoS.
>
> Now the question: is it possible to obtain the correct value of
> out-counters in iptables?
>
> Regards,
> D.

Where do you account these 10Mbytes per second? What tool do you use? What 
kind of shaping discipline you use? I suppose a simple TBF may sort you 
problems out. The bottleneck with those modems is their ethernet 
subinterface, which tends to provide 100Mbit/sec, but the modem can upload 
just a tiny bit of those. An egress classless qdics will fix your 
problems. TBF of pfifo_fast which is default.


-- 
Using M2, Opera's revolutionary e-mail client: http://www.opera.com/m2/

Re: Traffic counts wrongly in OUTPUT.

[deniska]

Daniel wrote:
> Where do you account these 10Mbytes per second? What tool do you use? What 

# iptables -I OUTPUT 1 -o ppp0 -j LOG
# iptables -Z OUTPUT; while true; do iptables -L -v | egrep 'OUTPUT|LOG'; sleep 1; done

Chain OUTPUT (policy ACCEPT 0 packets, 0 bytes)                                                                   0     0 LOG        all  --  any    ppp0    anywhere             anywhere            LOG level warning 
Chain OUTPUT (policy ACCEPT 11409 packets, 17M bytes)
11409   17M LOG        all  --  any    ppp0    anywhere             anywhere            LOG level warning 
Chain OUTPUT (policy ACCEPT 20800 packets, 30M bytes)
20800   30M LOG        all  --  any    ppp0    anywhere             anywhere            LOG level warning 
Chain OUTPUT (policy ACCEPT 30292 packets, 44M bytes)
30292   44M LOG        all  --  any    ppp0    anywhere             anywhere            LOG level warning 
Chain OUTPUT (policy ACCEPT 39745 packets, 58M bytes)
39745   58M LOG        all  --  any    ppp0    anywhere             anywhere            LOG level warning 
Chain OUTPUT (policy ACCEPT 49285 packets, 72M bytes)
49285   72M LOG        all  --  any    ppp0    anywhere             anywhere            LOG level warning 
Chain OUTPUT (policy ACCEPT 58688 packets, 85M bytes)
58688   85M LOG        all  --  any    ppp0    anywhere             anywhere            LOG level warning 
Chain OUTPUT (policy ACCEPT 68216 packets, 99M bytes)
68216   99M LOG        all  --  any    ppp0    anywhere             anywhere            LOG level warning 
Control-C

> kind of shaping discipline you use? I suppose a simple TBF may sort you 
> problems out. The bottleneck with those modems is their ethernet 
> subinterface, which tends to provide 100Mbit/sec, but the modem can upload 
> just a tiny bit of those. An egress classless qdics will fix your 
> problems. TBF of pfifo_fast which is default.

and my QoS setting...

tc qdisc add dev ppp0 root handle 1: prio
tc qdisc add dev ppp0 parent 1:2 tbf latency 50ms burst 1450 rate 50kbit
tc qdisc add dev ppp0 parent 1:1 pfifo limit 200
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip dport 80 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip dport 25 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip dport 110 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip dport 143 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip dport 22 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip dport 21 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip sport 21 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip sport 25 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip sport 80 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip dport 20 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 1 u32 match ip dport 53 0xffff flowid 1:1
tc filter add dev ppp0 protocol ip parent 1: prio 2 u32 match u32 0 0x0 flowid 1:2

Regards,
D.

Traffic counts wrongly in OUTPUT.

[deniska]

Hi,

My counter in OUTPUT grows with >10MBytes/sec but I have only 64kbit of
bandwidth. I have QoS shaping outgoing traffic, so, my guess is, the
kernel sends packets as quick as possible but most of them are
dropped in a queue by QoS.

Now the question: is it possible to obtain the correct value of
out-counters in iptables?

Regards,
D.