Re: [LARTC] Does anyone have a working proxyARP setup?
gypsy wrote: ... gypsy ifconfig eth0 x.x.x.96 broadcast x.x.x.111 netmask 255.255.255.240 gypsy ifconfig eth1 x.x.x.96 broadcast x.x.x.111 netmask 255.255.255.240 ... I think you can't use x.x.x.96 here, because it is the address of your network x.x.x.96/28. Useable ip addresses are .97 - .110. And you can't have the same ip address and netmask on two interfaces. Use maybe 'netmask 255.255.255.255' on one of them. As far as the question in the subject is concerned, yes, I have. -- Martin ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] HTB weird problem ....
if i remove the 1 packet ... it would be again exceed the ceiling ... thanks ill try r u referring to this faq in docum??!?!?! http://www.docum.org/docum.org/faq/cache/40.html the file linux/include/net/sched/pkt_sched.h does not have #define PSCHED_CLOCK_SOURCE PSCHED_CPU im using 2.4.20-22 kernel, maybel ill try 2.4.27 oh well i think i just have to check it out on monday ...thanks so much :D --- Andy Furniss [EMAIL PROTECTED] wrote: Drink Linux wrote: hello Andy , i think they are right for 256kbps = 2048kbit ... ahh I see. I just tried your setup on my eth0 and it works OK. Though HTB's stats don't seem too accurate - I used wget/ftp to judge rates. You may need to patch HTB/use a newer kernel - there was a patch posted on this list a while back which may affect you. Also you may need to set Hz higher or use psched = CPU for timing. See www.docum.org . i have added a leaf pfifo with a limit of 1 packet per second, coz if i have 2-10 it wont work...viola !!! the ceiling rate for each class rule is now working... my problem is that you can reach the ceiling class only if you have 4-5 files getting through FTP, ex: 256kbps Ceil 1 file ftp download = 80-90 kbps max speed 4-5 files ftp download = almost 256kbps how can i make it work to 256kbps speed for 1 file alone ...? Get rid of the 1 packet pfifo :-) Andy. ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/ ___ Do you Yahoo!? Declare Yourself - Register online to vote today! http://vote.yahoo.com ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] HTB weird problem ....
Drink Linux wrote: if i remove the 1 packet ... it would be again exceed the ceiling ... thanks ill try When you fix HTB you won't need it. r u referring to this faq in docum??!?!?! http://www.docum.org/docum.org/faq/cache/40.html the file linux/include/net/sched/pkt_sched.h include/net/pkt_sched.h is the one I changed on a 2.4.24. does not have #define PSCHED_CLOCK_SOURCE PSCHED_CPU im using 2.4.20-22 kernel, maybel ill try 2.4.27 2.4.27 should fix things HTB has been patched since 2.4.20. If you have 8 Mbit wirless your ceil/master rates need to be a bit less to allow for overheads. Andy. ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] Sending and receiving
On Saturday 09 October 2004 03:19, Alexis wrote: Hi all. Here's the situation Linux box with eth0 connected to LAN, and eth1 connected to internet via cablemodem. Connected to the lan are some voip devices, ive configured htb in eth1 to save some bandwith for the voip devices. Now i have another issue, at some hours of the days, some servers in the lan downloads data from other servers in internet and they use all bandwith available. My question is the following. Applying some classes to eth0 is a good way to reserve some bandwith for the traffic that comes from internet to the voip devices? Yes. I mean, is this a good way to manage the download traffic? Yes. Stef -- [EMAIL PROTECTED] Using Linux as bandwidth manager http://www.docum.org/ ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] Does anyone have a working proxyARP setup?
Martin Volf wrote: gypsy wrote: ... gypsy ifconfig eth0 x.x.x.96 broadcast x.x.x.111 netmask 255.255.255.240 gypsy ifconfig eth1 x.x.x.96 broadcast x.x.x.111 netmask 255.255.255.240 I think you can't use x.x.x.96 here, because it is the address of your network x.x.x.96/28. Useable ip addresses are .97 - .110. And you can't have the same ip address and netmask on two interfaces. Use maybe 'netmask 255.255.255.255' on one of them. -- Martin I have tried all IPs in the range, but I have not tried different netmasks. Thanks for that tip. Could you please post the output of 'route -n', 'ip route' and 'ip neigh show' as well as any 'ip route [add|del|*]' commands you run? I really believe that either the kernel thinks there are spoofed IPs or - most likely - that my routing table is junk. Here is a quote from http://www.sjdjweis.com/linux/proxyarp/ which is why I set both the same: After you have the above steps done, you will need to configure your network cards. This step should be done off of the network since you may end up with some conflicting addresses. Give two NIC's identical IP addresses, subnet masks, and gateways. The IP you choose needs to be an unused address on your network. In my case, I used x.x.x.98, since my router is at x.x.x.97. You could actually use about any address on the wire that isn't in use. gypsy ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] Ceiling question
On Saturday 09 October 2004 01:30, Peter Huetmannsberger wrote: Hi! I have a setup where I want to prefer traffic on one port (for testing purposes I used port 22) my setup is : tc qdisc add dev eth3 root handle 1: htb default 30 tc class add dev eth3 parent 1: classid 1:1 htb rate 96mbit burst 15k tc class add dev eth3 parent 1: classid 1:7 htb rate 2mbit burst 15k tc class add dev eth3 parent 1:1 classid 1:10 htb rate 96mbit burst 15k tc class add dev eth3 parent 1:7 classid 1:20 htb rate 1800kbit ceil 2mbit burst 15k tc class add dev eth3 parent 1:7 classid 1:30 htb rate 200kbit ceil 2mbit burst 15k The parent of class 1:7 should be 1:1. tc qdisc add dev eth3 parent 1:10 handle 10: sfq perturb 10 tc qdisc add dev eth3 parent 1:20 handle 20: sfq perturb 10 tc qdisc add dev eth3 parent 1:30 handle 30: sfq perturb 10 U32=tc filter add dev eth3 protocol ip parent 1:0 prio 1 u32 $U32 match ip src 81.223.175.128/26 flowid 1:10 $U32 match ip dst 192.168.5.9 match ip sport 22 0xfff flowid 1:20 $U32 match ip dst 192.168.5.9 match ip dport 22 0xfff flowid 1:20 $U32 match ip dst 192.168.5.10 match ip sport 22 0xfff flowid 1:20 $U32 match ip dst 192.168.5.10 match ip dport 22 0xfff flowid 1:20 What would like to achieve is that trafic on port 22 has 1800kbit always, regardless of traffic on any other port, but if there is no traffic on port 22 the rest can claim the whole bandwidth (i.e. 2.3 mbit ). However if I set the ceiling to 2mbit on both, they seem to sher the bandwidth evenly. Mhh, it should work. If I set the ceiling to 512k on 1:30, I get better performance on 1:20. Mhh, Do I not understand the concept correctly? I assumes that the rate would give me the guaranteed bandwidth for each class, Indeed. and the ceiling is there to make it use what's left over from the other classes. The ceil is the maximum the class can send. I did some tests, maybe they can help you to understand htb: http://www.docum.org/docum.org/tests/ Stef -- [EMAIL PROTECTED] Using Linux as bandwidth manager http://www.docum.org/ ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
[LARTC] weird problem with ip+snat+tun0
i have a box with 2 real interfaces and one more virtual eth0 - to the internet (193 eth1 - to the local net (192.168..) tun0 - to another ISP the routing is: all the free/local classes i send them directly on eth0, the rest of the internet i send throw tun0 the admin from tun0 wants me to snat all the packets with my end of the ip-tun0-interface and i snat all the trafic that go to local/free nets the problem is that on the tun0 i see packets with source adr my eth0 and dest somewhere in the internet, and are only acks (i also see nated trafic), why ill start with some confs and at the end some descoveryes: so a ip rule looks like: 0: from all lookup local 32516: from 192.168.40.0/24 lookup metro 32517: from 192.168.40.254 lookup tunel 32518: from 192.168.40.253 lookup tunel .. 32765: from 192.168.40.2 lookup tunel 32766: from all lookup main 32767: from all lookup default an ip route list table metro have entres like: 84...0/17 via 193. dev eth0 an ip route list table tunel its only a default default via 10.0.1.1 dev tun0 an the main have the directed connected nets and a def throw eth0 the iptables looks: filter - empty mangle - mark trafic for the tc part nat - only Chain POSTROUTING 481 52825 SNAT all -- * tun0 192.168.40.0/24 0.0.0.0/0 to:10.0.1.2 0 0 SNAT all -- * eth0 192.168.40.100 0.0.0.0/0 to:IP_IF_ETH0 a tcpdump on tun0 gets tcpdump -i tun0 -n | grep -v 10.0.1.2 IP_IF_ETH0.8181 24.129.71.219.42694: ack 2449728106 win 33870 (DF) IP_IF_ETH0.8181 24.129.71.219.42694: ack 1 win 33870 (DF) IP_IF_ETH0.8181 81.208.36.95.9195: . ack 272319646 win 65225 (DF) so i begin to put accounting/logging rules in iptables with -s IP_IF_ETH0, i did in nat POSTROUTING, in filter OUTPUT,INPUT,FORWARD, and i got on OUTPUT Oct 10 04:10:39 kernel: IN= OUT=eth0 SRC=IP_IF_ETH0 DST=83.175.129.103 LEN=40 TOS=0x00 PREC=0x00 TTL=64 ID=0 DF PROTO=TCP SPT=8181 DPT=4894 WINDOW=0 RES=0x00 ACK RST URGP=0 so its a localgenerated packet that is marked to get out on eth0, but he gets on tun0. I presumes (pls confirm) that the label of the interface is put by the output_routing, and when he gets to the OUTPUT_conntrack its marked to get out on tun0 but dont modify the label, so he dont match my rule of snat -o tun0 how can i solve the problem, i dont see how, or its the config bad, or a bug :-))) C ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/