Thank you Andy. I'll have another look at it and change the values to
get it to do what I need. Again, thank you.
Matthew Pearson
Andy Furniss wrote:
Matthew Pearson wrote:
I am using the script below to simulate a very low bandwidth
connection. I found that I could turn the bandwidth knob down to
about 4kbit, but below that I didn't get any traffic through. I've had
a look at this generally, but couldn't find an answer. It doesn't even
seem like the first reply packet gets through. I have tried it with
much bigger buffers, but this doesn't help.
I found that if I put a web proxy on the machine that is running this,
then the minimum I can turn the bandwidth down to is 12kbit and below
that the web browser doesn't get anything back.
Is this because the delay is so great that things are getting thrown
away by the kernel? Could I munge the packets to turn up the TTL or
something similar?
Many thanks for some excellent tools.
Matthew Pearson
#!/bin/bash
CLIENT1=192.168.1.190/32
CLIENT2=192.168.1.191/32
OPER=add;
DEV=eth0
RATE=3kbit
PEAKRATE=3kbit
BUFFER1=10kb
BUFFER2=10kb
echo -e "Attach Egress policy..."
tc qdisc $OPER dev $DEV root handle 1:0 htb default 15
tc class $OPER dev $DEV parent 1:0 classid 1:1 htb rate 240kbit
tc class $OPER dev $DEV parent 1:1 classid 1:2 htb rate 240kbit ceil
240kbit
tc class $OPER dev $DEV parent 1:1 classid 1:3 htb rate 240kbit ceil
240kbit
tc class $OPER dev $DEV parent 1:1 classid 1:15 htb rate 240kbit ceil
240kbit
tc qdisc $OPER dev $DEV parent 1:2 handle 2:0 tbf rate $RATE burst
$RATE limit $BUFFER1 peakrate $PEAKRATE mtu 1600
I don't really get using tbf under htb - but it may be OK.
The reason it fails <12kbit is because you use it for burst - which is a
buffer length so <12kbit won't pass a 1500 byte packet.
Andy.
_______________________________________________
LARTC mailing list
LARTC@xxxxxxxxxxxxxxx
http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
