Re: [LARTC] graphing
> Does anyone know what Martin used to generate the graphs in the HTB user > guide? > > http://luxik.cdi.cz/~devik/qos/htb/manual/userg.htm > > Thanks! > > -Jacob I think Martin used the following tool to transmit frames: http://luxik.cdi.cz/~devik/qos/ethloop/ and I suppose he used gnuplot to graph, I also did some graphs on HTB visible at http://rawsoft.org with ethloop and gnuplot. Vincent. > ___ > LARTC mailing list / LARTC@mailman.ds9a.nl > http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/ > ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] tail and head drop qdiscs
> vincent perrier wrote: > > I think that there are no qdiscs that permit to drop the oldest > > frame of a queue when this queue is full ... > What sort of thing do you want this for? > ... > Andy. > Very low rate (satellite links for example) and strict priority can lead to messages waiting for a chance to be transmitted and getting old in low prio queues as higher priority messages take all the bandwidth. Vincent. > ___ > LARTC mailing list / LARTC@mailman.ds9a.nl > http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/ > ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] graphing
On Tuesday 25 January 2005 20:18, jacob martinson wrote: > Does anyone know what Martin used to generate the graphs in the HTB user > guide? Yep, it's a perl script to process the counters and I gnuplot to create the graph. I used the same scripts to create my graphs, so I downloaded it somewhere from his site. If you can't find the scripts, let me know and I will try to find mine. Stef ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] graphing
On Tuesday 25 January 2005 20:18, jacob martinson wrote: > Does anyone know what Martin used to generate the graphs in the HTB user > guide? Do you mean the class tree graphs? You can use GraphViz to generate those, maybe with the help of a script like http://www.metamorpher.de/files/tc-graph.pl which creates graphs (together with GraphViz 'dot') like these http://www.metamorpher.de/files/fairnat.png (big!). If you mean the statistic graphs, maybe these were done with RRDTool or similar. There are scripts that create this kind of graph as well, but I haven't got an example URL handy. Maybe someone else on this list can provide one. HTH Andreas ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
[LARTC] graphing
Does anyone know what Martin used to generate the graphs in the HTB user guide? http://luxik.cdi.cz/~devik/qos/htb/manual/userg.htm Thanks! -Jacob ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] Htb, imq and sfq traffic shaping
Ixen Gerthannes wrote: Hello, I wrote a simple traffic shaping script (below) that should have allowed me to shape my internet traffic a bit (ppp0 - adsl 128kbit/64kbit; local interface eth0). The script works only partially - the speed is being limited but too much. Without running this script my download rate is about 10kBytes (with second computer also downloading at about 6kBytes). After running it my download speed decreases to about 1..2kBytes and so the second computer. Could you tell me what have I done wrong ? Thank you for replies. Ixen -- My configuration: If you are shaping on and using PC1 and forwarding traffic for PC2 then you only need one IMQ. You need to send all inbound on ppp0 to it in PREROUTING. Shape egress directly on ppp0. Traffic from the shaping PC will have ppp0s address not eth0s, you need to mark local addresses in postrouting mangle for the queues on ppp0. IMQ needs to be set to hook after nat in prerouting. This may be the default AB? To filter inbound local (denatted) ip addresses you can't mark with iptables you need to use tc filters. Remember traffic to/from the shaping PC over ppp0 will have ppp0s address. You need to back off from 128/64 - maybe 100/50, it depends what you want to shape for. There are other tweaks if you really care about latency. You need to sacrifice downstream bandwidth - and upstream unless you are prepared to find link overheads and tweak/patch. Andy. ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] Re: Confuse, putting packets in wrong mangle table.
Rio Martin. wrote: On Wed, 19 Jan 2005 09:19:58 +0100 Tóth Nándor <[EMAIL PROTECTED]> wrote: Rio Martin. wrote: Folks, I am a little bit confuse in how to put these packets into correct mangle table for traffic shaping. This is what i ve planned to do: - - - - :eth0 [ LINUX-BOX ] eth1: - - - - Let say: eth0: 220.100.1.1 eth1: 192.168.1.1 eth1:1 192.168.1.2 192.168.1.0/24 get natted into 220.100.1.1 before reaching the internet. I put every packets coming from internet (eth0) this way: # iptables -t mangle -I PREROUTING -i eth0 .. * * * * But what if i would like to do the same way with ip 192.168.1.1 and 192.168.1.2 ? How to put the packets in correct mangle table? You can only shape traffic leaving your interface. You have to shape the incoming traffic at interface eth1, the outgoing traffic at interface eth0. This site is great for learning: http://www.knowplace.org/shaper/qdisc.html#egress I found to use iptables+CLASSIFY method the easiest way to classify packets. Your question is a little dizzy... Oopss sorry then :)) Actually, there is squid proxy running in my box, with those two private IPs (192.168.1.1 & 192.168.1.2) Both get natted to 220.100.1.1. Is there only one proxy running? I need to shape incoming traffic to both of these ips but i am affraid i have to face that i am not able to shape traffic which is generate from this box unless those two IPs were outside the box. Maybe true - maybe not you would need to test with imq. There is also a kernel option to do with nat of local connections. If i have one more public IP than i should not so much worry about, cause i can shape it using IMQ. So any other ideas maybe? i'm thinkin just in case IMQ would help .. Possibly - but I don't understand your setup or what you want to shape :-) Thanks for the info u gave anyway.. - Rio.Martin - ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/ ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] tail and head drop qdiscs
vincent perrier wrote: I think that there are no qdiscs that permit to drop the oldest frame of a queue when this queue is full, but I would like to be wrong: bfifo drops arriving frames when the max queue length is reached. red also drops arriving frames in a more elaborate fashion, with a drop probability that increases above a limit and becomes a drop certitude when the max queue length is reached. sfq drops "fairly", it piles frames with same (ip source, ip dest, port source and port dest) and drops a frame from the biggest pile to favour the smaller flows. But I cannot find the qdisc that reacts like a bfifo or red qdisc but drops the oldest message of the queue when a drop is needed. Even better: a qdisc with an autodestruct algorithm for each frame in the queue that would garantee the "freshness" of each frame coming out. If you know of such qdiscs, please share. What sort of thing do you want this for? I made a version of esfq head drop for use in specific circumstances. The only other thing I thought head drop generally could be usefull for is a game server, but then maybe a policer would be better than a queue WRT latency. Andy. ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] tc filter change/replace
Tomas Graf: You mentioned that "All classifier except rsvp are changeable in the latest bk. The patchset has gone in around rc1-bk3-5" Well issue 1 is that in iproute-latest (as well as the one released yesterday) the output of: tc show filter dev eth0 parent :" and change/replace DO NOT WORK. --- filter protocol ip pref 50 u32 fh 800::801 order 2049 ht 800 bkt 0 match 0a010102/ at 12 police 0x4 rate 2bit burst 2559b mtu 2kb action reclassify ref -1035812 bind -897856 on the bk version (latest changeLog entry 18-1-2005) as you suggested the output is NOT VERBOSE ENOUGH to see if all parameters are working, but change/replace SEEM TO BE WORKING. Heres the output: filter protocol ip pref 50 u32 fh 800::801 order 2049 ht 800 bkt 0 match 0a010102/ at 12 ISSUE 2 --- In lots of computers i try to install iproute2 i get an error about db_185.h missing . Issuing a simple apt-get install libdb4.1-dev resolves the problem and the arpd compiles perfectly. Maybe consider documenting that somewhere. --- Thomas Graf <[EMAIL PROTECTED]> wrote: > * Antonios Chalkiopoulos > <[EMAIL PROTECTED]> > 2005-01-24 14:02 > > > > I am wondering whether the commands 'tc filter > change' and 'tc filter replace' > > are working. Googling around revealed a few emails > on the LARTC stating that > > unfoirtunately they do not work. A quick check > showed that they are still not > > implemented (properly). > > Some propose a solution to build a second tree of > filters and apply the 2nd > > one when a need for it arises. Howver how can > someone achive the above trick. > > All classifiers except rsvp are changeable in the > latest bk. The > patchset has gone in around rc1-bk3-5. > ___ > LARTC mailing list / LARTC@mailman.ds9a.nl > http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: > http://lartc.org/ > __ Do you Yahoo!? Take Yahoo! Mail with you! Get it on your mobile phone. http://mobile.yahoo.com/maildemo ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
Re: [LARTC] Info about VoiP Qos
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Ed Wildgoose wrote: | Please reply onlist in general | Sorry, i didn't notice the answer went to you only.. i did the same mistake with Jason. | Check the HTB code in iproute and look for the "hysteris" variable. | Basically it just lets a couple of packets burst out at the same time | instead of doinga very granular per packet throttling. ok.. | Find out what the protocol is on the line. ADSL for example uses ATM to | encapsulate each packet. YOu buy 512Kbs of ATM bandwidth, which | actually works out as somewhat less TCP bandwidth (which is what you are | throttling). So do some speed tests on the line and work out the REAL | speed of the line. Preferably use small packets since the overhead of | encapsulation often shoots up then. | | This might not be relevant though - just a warning though I understand it's a frame relay | I meant, reserve (say) 10Kbs, and only share the rest... I see.. this way it works. Things are getting a little more complicated though.. I tested with some people, calling ip to ip, and it works just fine. I understand that, if it's just a matter of my bandwith being overloaded and not correctly shaped i should get the same disturbs, but i do not. I suspect there is something involved on the server that converts the call from ip to POTS and vice versa too.. Thanks! Ricky -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.4 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFB9hXNwfVi6Z6cQK8RAoMNAJ9p8OJ1amuS+y0VKVla/CTZRwuwugCfT3xK rMU221VkLji/lhh3t4xXbKk= =JZ5A -END PGP SIGNATURE- ___ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/