Re: [LARTC] any errors in this script?

[Stef Coene <stef.coene@docum.org>]

On Thursday 07 August 2003 14:08, Tom Verbeek wrote:
> Hi,
>
> ok, I dropped the idea of using Wondershaper without having background
> knowledge and are now trying to set up my own script :)
ok :)

> my max. upload bandwidth is 192 Kbit/s and my goal is to ensure smooth
> gameplay (udp) while allowing other protocols to use up the remaining
> bandwidth. Mail is supposed to be treated with the least priority.
>
> Cheers,
> Tom.
>
> #-----------------------------------------------------
> # clean existing down- and uplink qdiscs, hide errors
> tc qdisc del dev ppp0 root    2> /dev/null > /dev/null
> tc qdisc del dev ppp0 ingress 2> /dev/null > /dev/null
>
> #root htb with non-classified data to 1:11
> tc qdisc add dev ppp0 root handle 1: htb default 11
>
> #max uplink class
> tc class add dev ppp0 parent 1: classid 1:1 htb rate 180kbit ceil 180kbit
> burst 2k
>
> #interactice class
> tc class add dev ppp0 parent 1:1 classid 1:10 htb rate 130kbit ceil 180kbit
> burst 2k prio 0
>
> #default class
> tc class add dev ppp0 parent 1:1 classid 1:11 htb rate 40kbit ceil 180kbit
> burst 2k prio 1
>
> #slow class
> tc class add dev ppp0 parent 1:1 classid 1:12 htb rate 10kbit ceil 70kbit
> burst 2k prio 2
>
> #put udp into interactive class
> tc filter add dev ppp0 parent 1:0 protocol ip prio 0 u32 \
>    match ip protocol 17 0xff \
>    flowid 1:10
>
> #put mail into slow class
> tc filter add dev ppp0 parent 1:0 protocol ip prio 2 u32 \
>    match ip sport 25 0xffff \
>    flowid 1:12
>
> tc filter add dev ppp0 parent 1:0 protocol ip prio 2 u32 \
>    match ip dport 25 0xffff \
>    flowid 1:12
>
> tc filter add dev ppp0 parent 1:0 protocol ip prio 2 u32 \
>    match ip sport 110 0xffff \
>    flowid 1:12
>
> tc filter add dev ppp0 parent 1:0 protocol ip prio 2 u32 \
>    match ip dport 110 0xffff \
>    flowid 1:12

You set a burst of 2k and that's too low.  Just remove the burst parameter so 
htb can calculate the minimal burst for you.

Different prio's for the filters makes no different because it's only used to 
order the filters.

DIfferent prio's for htb classes can be dangerous.  If a low prio class like 
class 1:10 receives more data then it's rate (130kbit), the latency can go 
way up.  So it's important the you monitor your setup so your low prio 
classes are never overlimited.

Stef

-- 

stef.coene@docum.org
 "Using Linux as bandwidth manager"
     http://www.docum.org/
     #lartc @ irc.oftc.net

_______________________________________________
LARTC mailing list / LARTC@mailman.ds9a.nl
http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/