[LARTC] Difference between tc reported rate and IPTraf Rate
Iam using HTB+IMQ on an Ethernet bridge and tc for statistics reporting. I am finding that tcreportedrateisabout 5-10% less than that reported by IPTraf for the same traffic. Has anyone observed this before? Which is more accurate? Rajesh
[LARTC] Difference between tc reported rate and IPTraf Rate
Title: [LARTC] Difference between tc reported rate and IPTraf Rate Trevor, Thanks for pointing out one of the possible reasons for the difference. However, in this case I have also tried using the local display and got similar results ( though the difference reduces a bit). Thanks, Rajesh Hey pal, What about the overhead part to create the display that flows betweenthe internet gateway(Machine you are monitoring) and your machine???.TrevorOn Mon, 2003-07-14 at 14:35, Rajesh Srivastava wrote: I am using HTB+IMQ on an Ethernet bridge and tc for statistics reporting. I am finding that tc reported rate is about 5-10% less than that reported by IPTraf for the same traffic. Has anyone observed this before? Which is more accurate? Rajesh -- ( - LINUX, It's all about CHOICE - )/~\ __ [EMAIL PROTECTED] __ /~\| \) / Pre Sales Consultant - Red Hat \ (/ ||_|_ \ 9820349221(M) | 22881326(O) / _|_| \___/ Previous message: [LARTC] Difference between tc reported rate and IPTraf Rate Next message: [Re: [LARTC] marking in OUTPUT --mangle; locally generated packets and route lookup - broken?] Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
[LARTC] Combining ingress and egress ( IMQ+HTB)
I am successfully running ingress (IMQ) and egress (HTB) shaping on a bridge. Is there any way to combine and share the bandwidth between ingress and egress? Example: I have set up www service for egress at 128 KB and ingress at 256 KB. The shaping on them works fine separately. However, I want to create a single virtual pipe for www traffic and limit both ingress and egress combined to 256 KB. Thanks. Rajesh ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Fw: [LARTC] Combining ingress and egress ( IMQ+HTB)
[LARTC] Combining ingress and egress ( IMQ+HTB)I checked the guide and you are right about it. However, I cannot get it to work, if I remove egress qdisc the traffic stops following any rules. Has anyone done this successfully before? Any pointers are welcome. Thanks Rajesh The way I understand it is summed up in a quote from LEAF Bering user's guide --- In many cases like those of ISPs, the bandwidth allocation is for incoming and outgoing combined. Under such situations, in stock linux, a virtual device called IMQ has been created through which all traffic passes. Thus shaping on IMQ will enable shaping total traffic and not incoming and outgoing separately. --- On Wednesday June 25 2003 02:37 pm, Rajesh Srivastava wrote: I am successfully running ingress (IMQ) and egress (HTB) shaping on a bridge. Is there any way to combine and share the bandwidth between ingress and egress? Example: I have set up www service for egress at 128 KB and ingress at 256 KB. The shaping on them works fine separately. However, I want to create a single virtual pipe for www traffic and limit both ingress and egress combined to 256 KB. Thanks. Rajesh ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/ -- Regards Joseph Watson ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] tc show error for ingress
Stef, tc -s -d class show dev eth0 parent : gives no output. I am using the latest tc which has been download and built with the 2.4.20 kernel includes. Thanks, Rajesh On Friday 06 June 2003 16:39, Rajesh Srivastava wrote: Hi, I am trying to do both ingress and egress bandwidth management on a bridge. eth0 is the WAN interface. IPTraf shows that the following script is running successfully and it limits bandwidth both ways to 256 KB However when I want to see the statistics using tc - it only shows me stats for htb class and not for ingress class. The tc output is as follows: tc -s -d class show dev eth0 class htb 1:1 root prio 0 quantum 3276 rate 256Kbit ceil 256Kbit burst 1926b/8 mpu 0b cburst 1926b/8 mpu 0b level 0 Sent 193313679 bytes 189055 pkts (dropped 1, overlimits 0) rate 32656bps 32pps backlog 9p lended: 189046 borrowed: 0 giants: 0 tokens: -77245 ctokens: -77245 What is wrong here? Try tc -s -d class show dev eth0 parent : Stef -- [EMAIL PROTECTED] Using Linux as bandwidth manager http://www.docum.org/ #lartc @ irc.oftc.net ___ LARTC mailing list / [EMAIL PROTECTED] http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
[LARTC] fwmark on bridge+htb
Hi,Fw marking using IP Tables does not work on an Ethernet bridge. Is itpossible to set up the mark using iptables so that the packet can bedirectly routed to a predefined class using a mark value?The advantage of this would be the one can use conntrack etc. to trackconnections and forward the packets to the correct class.Example:tc qdisc del dev eth0 roottc qdisc add dev eth0 root handle 1: htb default 10tc class add dev eth0 parent 1: classid 1:1 htb rate 512kbit ceil 512kbittc class add dev eth0 parent 1:1 classid 1:10 htb rate 128kbit ceil 256kbittc class add dev eth0 parent 1:1 classid 1:22 htb rate 64kbit ceil 256kbitprio 3tc class add dev eth0 parent 1:1 classid 1:80 htb rate 64kbit ceil 128kbitprio 3# traditional method of classifying traffic into flowidstc filter add dev eth0 parent 1:1 protocol ip prio 3 u32 match ip sport 220x flowid 1:22tc filter add dev eth0 parent 1:3 protocol ip prio 3 u32 match ip sport 800x flowid 1:80---What I want to achieve is as follows#set predefined marksiptables -t mangle -A PREROUTING -i eth0 -p tcp --sport 22 -m state --stateESTABLISHED -j MARK --set-mark 22iptables -t mangle -A PREROUTING -i eth0 -p tcp --sport 80 -m state --stateESTABLISHED -j MARK --set-mark 80Now I want to be able to forward packets marked 22 to class 1:22 and thosemarked 80 sent to class 1:80 without using the tc fw filter ( as it does notwork on bridges ).Any help or pointers shall be highly appreciated.ThanksRajesh
[LARTC] fw mark on Ethernet bridge + HTB
Hi, Fw marking using IP Tables does not work on an Ethernet bridge. Is it possible to set up the mark using iptables so that the packet can be directly routed toa predefinedclass using a mark value? The advantage of this would be the one can use conntrack etc. to trackconnections and forward the packets to the correct class. Example: tcqdisc del dev eth0 root tc qdisc add dev eth0 root handle 1: htb default 10 tcclass add dev eth0 parent 1: classid 1:1 htb rate 512kbit ceil 512kbit tcclass add dev eth0 parent 1:1 classid 1:10 htb rate 128kbit ceil 256kbit tcclass add dev eth0 parent 1:1 classid 1:22 htb rate 64kbit ceil 256kbit prio 3 tc class add dev eth0 parent 1:1 classid 1:80 htb rate 64kbit ceil 128kbit prio 3 # traditional method of classifying traffic into flowids tcfilter add dev eth0 parent 1:1 protocol ip prio 3 u32 match ipsport22 0x flowid 1:22 tcfilter add dev eth0 parent 1:3 protocol ip prio 3 u32 match ip sport80 0x flowid 1:80 --- What I want to achieve is as follows #set predefined marksiptables -t mangle -A PREROUTING -ieth0 -p tcp--sport 22 -m state --state ESTABLISHED -j MARK --set-mark22iptables -t mangle -A PREROUTING -ieth0 -p tcp --sport 80 -m state --state ESTABLISHED -j MARK --set-mark 80 Now I want to be able to forward packets marked 22 to class 1:22 and those marked 80 sent to class 1:80 without using the tc fw filter ( as it does not work on bridges ). Any help or pointers shall be highly appreciated. Thanks Rajesh
