This script only really helps for egress shaping for those who have DSL lines and a lot of uplink traffic, and want to bring down response times for gaming etc...
Try FairNat , It doesn't support multiple subnets, but since you have seperate external IP's, it might just be able to help. I use(and modified mine), and currently working very well. Thinking of trying HSFC , instead of HTB in the future, but getting docs that make sense is my current problem...
I hope this helps, since I am also quite new to all this myself. But sure is fun fiddling ;-).
Jaques
On 21/11/06, drew einhorn <
drew.einhorn@xxxxxxxxx> wrote:
I'm trying to shape traffic on a Devil-Linux box.
This note was originally sent to their maillist,
because the LARTC list appears to have been down
for the past few days. My mailbox was just flooded
with a half dozen or so confirmation requests in response
to my repeated attempts to subscribe to this list.
---------- Forwarded message ----------
From: drew einhorn <drew.einhorn@xxxxxxxxx>
Date: Nov 19, 2006 11:51 PM
Subject: Traffic Shaping on a Transparent Bridge not working!
To: devil-linux-discuss@xxxxxxxxxxxxxxxxxxxxx
My first DL project was going well. Then I ran into problems attempting
to shape my bandwidth.
First I'll describe the parts that I believe are working correctly.
I have a DL 1.2.11 box running the default kernel, 2.4.33.3-grsec
I have br0 bridging all four ports eth0, eth1, eth2, eth3 on a quad port
pci card. The bridge has not been assigned an ip number on the theory
that this makes it much more difficult to attack. The bridge connects
four devices on the 3bit public static ip block from my ISP.
I have a single port ethernet pci card, eth4 with a static ip, on my
internal private ip network. It is used for remote managent of the DL
box from anywhere on my internal network.
eth0 is connected to my ISP's router via the ethernet port on my
ISDN modem. I know ISDN is a nearly dead technology, but it's the best
thing my crappy telco offers. Tried a satellite ISP, but that's another
long and sad story.
eth1 is connected to a hardened publicly accessible host.
eth2 and eth3 are connected to the WAN ports on a couple of Linksys
Cable/DSL routers. Eventually most of their functions will migrate to the
DL box, but that is more than I wanted to bite off in my first DL project.
The first Linksys box NATs one of my public ips to my internal private
ip network. The second Linksys box is newer and includes a wireless
access point used by a couple neighbors. It NATs a second public ip to
a separate private ip network.
All of the above appears to be working as expected.
After pondering the mysteries of traffic shaping I decided to start with
wondershaper 1.1a from lartc.org, rather than starting from scratch.
Tried both the cbq and htb versions without any success.
RTFM time. The htb section of http://lartc.org/howto/index.html is easier
reading than the cbq section. And the howto claims htb is better anyway.
Let's focus on the htb version of wondershaper.
OK, First we edit wshaper.htb and configure the shell variables. Then we
run: sh -x wshaper.htb
to echo the commands as they are executed.
Then we start pinging the router at the other end of the ISDN line.
Then we start downloading a file to generate some traffic that really
needs to be shaped.
Then we run: sh -x wshaper.htb status
to gather some statistics
then we kill the download.
then we sh -x wshaper.htb stop to shut down the malfunctioning shaper.
Here's the output from the ping:
$ ping 67.0.192.10
PING 67.0.192.10 ( 67.0.192.10) 56(84) bytes of data.
Link is idle, normal ping times.
64 bytes from 67.0.192.10: icmp_seq=0 ttl=254 time=48.5 ms
64 bytes from 67.0.192.10: icmp_seq=1 ttl=254 time=48.4 ms
64 bytes from 67.0.192.10: icmp_seq=2 ttl=254 time=48.4 ms
64 bytes from 67.0.192.10: icmp_seq=3 ttl=254 time= 48.4 ms
64 bytes from 67.0.192.10: icmp_seq=4 ttl=254 time= 48.5 ms
64 bytes from 67.0.192.10: icmp_seq=5 ttl=254 time=67.8 ms
64 bytes from 67.0.192.10: icmp_seq=6 ttl=254 time=48.3 ms
64 bytes from 67.0.192.10: icmp_seq=7 ttl=254 time=48.2 ms
Download starts. Shaping is not working! Queues in
router and/or ISDN modem grow, and ping times rapidly
become huge.
64 bytes from 67.0.192.10: icmp_seq=8 ttl=254 time=184 ms
64 bytes from 67.0.192.10: icmp_seq=9 ttl=254 time=1080 ms
64 bytes from 67.0.192.10: icmp_seq=10 ttl=254 time=2025 ms
64 bytes from 67.0.192.10: icmp_seq=11 ttl=254 time=1551 ms
64 bytes from 67.0.192.10: icmp_seq=12 ttl=254 time=1078 ms
64 bytes from 67.0.192.10: icmp_seq=13 ttl=254 time=896 ms
64 bytes from 67.0.192.10: icmp_seq=14 ttl=254 time=1088 ms
64 bytes from 67.0.192.10: icmp_seq=15 ttl=254 time=1171 ms
64 bytes from 67.0.192.10: icmp_seq=16 ttl=254 time=1272 ms
64 bytes from 67.0.192.10: icmp_seq=17 ttl=254 time=1280 ms
64 bytes from 67.0.192.10: icmp_seq=18 ttl=254 time=1101 ms
64 bytes from 67.0.192.10: icmp_seq=19 ttl=254 time=1258 ms
64 bytes from 67.0.192.10: icmp_seq=20 ttl=254 time=1211 ms
64 bytes from 67.0.192.10: icmp_seq=21 ttl=254 time=1259 ms
64 bytes from 67.0.192.10: icmp_seq=22 ttl=254 time=1373 ms
64 bytes from 67.0.192.10: icmp_seq=23 ttl=254 time=1424 ms
64 bytes from 67.0.192.10: icmp_seq=24 ttl=254 time=1461 ms
64 bytes from 67.0.192.10: icmp_seq=25 ttl=254 time=1277 ms
64 bytes from 67.0.192.10: icmp_seq=26 ttl=254 time=1521 ms
64 bytes from 67.0.192.10: icmp_seq=27 ttl=254 time=1467 ms
64 bytes from 67.0.192.10: icmp_seq=28 ttl=254 time=1335 ms
64 bytes from 67.0.192.10: icmp_seq=29 ttl=254 time=1329 ms
64 bytes from 67.0.192.10: icmp_seq=30 ttl=254 time=1386 ms
64 bytes from 67.0.192.10: icmp_seq=31 ttl=254 time=1360 ms
64 bytes from 67.0.192.10: icmp_seq=32 ttl=254 time=1416 ms
64 bytes from 67.0.192.10: icmp_seq=33 ttl=254 time=1480 ms
64 bytes from 67.0.192.10: icmp_seq=34 ttl=254 time=1345 ms
64 bytes from 67.0.192.10: icmp_seq=35 ttl=254 time=1356 ms
64 bytes from 67.0.192.10: icmp_seq=36 ttl=254 time=1370 ms
64 bytes from 67.0.192.10: icmp_seq=37 ttl=254 time=1278 ms
64 bytes from 67.0.192.10: icmp_seq=38 ttl=254 time=1612 ms
64 bytes from 67.0.192.10: icmp_seq=39 ttl=254 time=1520 ms
64 bytes from 67.0.192.10: icmp_seq=40 ttl=254 time=1322 ms
64 bytes from 67.0.192.10: icmp_seq=41 ttl=254 time=1545 ms
Kill the download. queues drain and ping times return to normal
64 bytes from 67.0.192.10 : icmp_seq=42 ttl=254 time=975 ms
64 bytes from 67.0.192.10: icmp_seq=43 ttl=254 time=67.4 ms
64 bytes from 67.0.192.10: icmp_seq=44 ttl=254 time= 73.6 ms
64 bytes from 67.0.192.10: icmp_seq=45 ttl=254 time=45.2 ms
64 bytes from 67.0.192.10: icmp_seq=46 ttl=254 time= 45.2 ms
64 bytes from 67.0.192.10: icmp_seq=47 ttl=254 time=44.8 ms
And, here's the shell commands and their output:
root@Devil:~ # sh -x wshaper.htb
+ DOWNLINK=100
+ UPLINK=100
+ DEV=eth0
+ NOPRIOHOSTSRC=
+ NOPRIOHOSTDST=
+ NOPRIOPORTSRC=
+ NOPRIOPORTDST=
+ '[' '' = status ']'
+ tc qdisc del dev eth0 root
+ tc qdisc del dev eth0 ingress
+ '[' '' = stop ']'
+ tc qdisc add dev eth0 root handle 1: htb default 20
+ tc class add dev eth0 parent 1: classid 1:1 htb rate 100kbit burst 6k
+ tc class add dev eth0 parent 1:1 classid 1:10 htb rate 100kbit burst 6k prio 1
+ tc class add dev eth0 parent 1:1 classid 1:20 htb rate 90kbit burst 6k prio 2
+ tc class add dev eth0 parent 1:1 classid 1:30 htb rate 80kbit burst 6k prio 2
+ tc qdisc add dev eth0 parent 1:10 handle 10: sfq perturb 10
+ tc qdisc add dev eth0 parent 1:20 handle 20: sfq perturb 10
+ tc qdisc add dev eth0 parent 1:30 handle 30: sfq perturb 10
+ tc filter add dev eth0 parent 1:0 protocol ip prio 10 u32 match ip
tos 0x10 0xff flowid 1:10
+ tc filter add dev eth0 parent 1:0 protocol ip prio 10 u32 match ip
protocol 1 0xff flowid 1:10
+ tc filter add dev eth0 parent 1: protocol ip prio 10 u32 match ip
protocol 6 0xff match u8 0x05 0x0f at 0 match u16 0x0000 0xffc0 at 2
match u8 0x10 0xff at 33 flowid 1:10
+ tc filter add dev eth0 parent 1: protocol ip prio 18 u32 match ip
dst 0.0.0.0/0 flowid 1:20
+ tc qdisc add dev eth0 handle ffff: ingress
+ tc filter add dev eth0 parent ffff: protocol ip prio 50 u32 match ip
src 0.0.0.0/0 police rate 100kbit burst 10k drop flowid :1
root@Devil:~ # sh -x wshaper.htb status
+ DOWNLINK=100
+ UPLINK=100
+ DEV=eth0
+ NOPRIOHOSTSRC=
+ NOPRIOHOSTDST=
+ NOPRIOPORTSRC=
+ NOPRIOPORTDST=
+ '[' status = status ']'
+ tc -s qdisc ls dev eth0
qdisc htb 1: r2q 10 default 20 direct_packets_stat 0
Sent 18649 bytes 191 pkts (dropped 0, overlimits 0)
qdisc sfq 10: parent 1:10 limit 128p quantum 1514b perturb 10sec
Sent 10582 bytes 147 pkts (dropped 0, overlimits 0)
qdisc sfq 20: parent 1:20 limit 128p quantum 1514b perturb 10sec
Sent 8067 bytes 44 pkts (dropped 0, overlimits 0)
qdisc sfq 30: parent 1:30 limit 128p quantum 1514b perturb 10sec
Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
qdisc ingress ffff: ----------------
Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
+ tc -s class ls dev eth0
class htb 1:1 root rate 100000bit ceil 100000bit burst 6Kb cburst 1724b
Sent 18649 bytes 191 pkts (dropped 0, overlimits 0)
rate 1320bit 1pps
lended: 0 borrowed: 0 giants: 0
tokens: 398459 ctokens: 108855
class htb 1:10 parent 1:1 leaf 10: prio 1 rate 100000bit ceil
100000bit burst 6Kb cburst 1724b
Sent 10582 bytes 147 pkts (dropped 0, overlimits 0)
rate 656bit 1pps
lended: 147 borrowed: 0 giants: 0
tokens: 398459 ctokens: 108855
class htb 1:20 parent 1:1 leaf 20: prio 2 rate 90000bit ceil 90000bit
burst 6Kb cburst 1711b
Sent 8067 bytes 44 pkts (dropped 0, overlimits 0)
rate 712bit
lended: 44 borrowed: 0 giants: 0
tokens: 432284 ctokens: 109555
class htb 1:30 parent 1:1 leaf 30: prio 2 rate 80000bit ceil 80000bit
burst 6Kb cburst 1699b
Sent 0 bytes 0 pkts (dropped 0, overlimits 0)
lended: 0 borrowed: 0 giants: 0
tokens: 503316 ctokens: 139264
+ exit
root@Devil:~ # sh -x wshaper.htb stop
+ DOWNLINK=100
+ UPLINK=100
+ DEV=eth0
+ NOPRIOHOSTSRC=
+ NOPRIOHOSTDST=
+ NOPRIOPORTSRC=
+ NOPRIOPORTDST=
+ '[' stop = status ']'
+ tc qdisc del dev eth0 root
+ tc qdisc del dev eth0 ingress
+ '[' stop = stop ']'
+ exit
root@Devil :~ #
Don't think we generated enough uplink traffic to exercise the htb qdiscs.
But it doesn't look like the ingress qdisc is working at all.
I'm out of ideas for now.
--
Drew Einhorn
_______________________________________________
LARTC mailing list
LARTC@xxxxxxxxxxxxxxx
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
_______________________________________________ LARTC mailing list LARTC@xxxxxxxxxxxxxxx http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
