On Sat, Apr 23, 2016 at 09:31:17PM -0400, Weedy wrote:
> On Sat, Apr 23, 2016 at 2:26 PM, Karl O. Pinc <[email protected]> wrote:
> > On Sat, 23 Apr 2016 12:32:50 -0400
> > Weedy <[email protected]> wrote:
> >
> >> For me qos-scripts used to be amazing. but ever since the fq_codel
> >> changes it just doesn't work the same.
> >> I mean how can a single http stream kill the network?
> >
> > I always begin by blaming buffer-bloat.  But that's just me
> > looking for something to blame.
> 
>     option upload        500
>     option download        10000
> Literally HALF of my line sync.

Hello Weedy,

Maybe you need some extra configuration for your line. Like, for a DSL
connection I add the linklayer adsl option and set an overhead value
like this:

Upload:
tc qdisc add stab overhead 40 linklayer adsl dev pppoe-wan root handle 1: hfsc 
default 40
Download:
tc qdisc add stab overhead 40 linklayer adsl dev ifb0 root handle 1: hfsc 
default 40

The overhead depends on the technology in use. There's some
documentation on this on the Internet.

Without this I also had to go way below my line speed to get halfway
decent latencies.

I don't know if you can get qos-scripts to set this for you. I use my
own script to configure the qdiscs.

Regards,
Sebastian

> 
> PING 4.2.2.1 (4.2.2.1) 56(84) bytes of data.
> 64 bytes from 4.2.2.1: icmp_seq=1 ttl=59 time=40.9 ms
> 64 bytes from 4.2.2.1: icmp_seq=2 ttl=59 time=40.4 ms
> 64 bytes from 4.2.2.1: icmp_seq=3 ttl=59 time=43.0 ms
> 64 bytes from 4.2.2.1: icmp_seq=4 ttl=59 time=38.9 ms
> 64 bytes from 4.2.2.1: icmp_seq=5 ttl=59 time=42.4 ms
> 64 bytes from 4.2.2.1: icmp_seq=6 ttl=59 time=39.8 ms
> 64 bytes from 4.2.2.1: icmp_seq=7 ttl=59 time=41.9 ms
> 64 bytes from 4.2.2.1: icmp_seq=8 ttl=59 time=42.0 ms
> 64 bytes from 4.2.2.1: icmp_seq=9 ttl=59 time=44.5 ms
> 64 bytes from 4.2.2.1: icmp_seq=10 ttl=59 time=44.0 ms
> 64 bytes from 4.2.2.1: icmp_seq=11 ttl=59 time=43.5 ms
> 64 bytes from 4.2.2.1: icmp_seq=12 ttl=59 time=46.7 ms
> 64 bytes from 4.2.2.1: icmp_seq=13 ttl=59 time=43.6 ms
> 64 bytes from 4.2.2.1: icmp_seq=14 ttl=59 time=43.0 ms
> 64 bytes from 4.2.2.1: icmp_seq=15 ttl=59 time=43.4 ms
> wget
> 64 bytes from 4.2.2.1: icmp_seq=17 ttl=59 time=78.9 ms
> 64 bytes from 4.2.2.1: icmp_seq=18 ttl=59 time=84.9 ms
> 64 bytes from 4.2.2.1: icmp_seq=19 ttl=59 time=89.3 ms
> 64 bytes from 4.2.2.1: icmp_seq=20 ttl=59 time=94.8 ms
> 64 bytes from 4.2.2.1: icmp_seq=21 ttl=59 time=100 ms
> 64 bytes from 4.2.2.1: icmp_seq=22 ttl=59 time=127 ms
> 64 bytes from 4.2.2.1: icmp_seq=23 ttl=59 time=143 ms
> 64 bytes from 4.2.2.1: icmp_seq=24 ttl=59 time=140 ms
> 64 bytes from 4.2.2.1: icmp_seq=25 ttl=59 time=104 ms
> 64 bytes from 4.2.2.1: icmp_seq=26 ttl=59 time=112 ms
> 64 bytes from 4.2.2.1: icmp_seq=28 ttl=59 time=128 ms
> 64 bytes from 4.2.2.1: icmp_seq=29 ttl=59 time=182 ms
> _______________________________________________
> openwrt-users mailing list
> [email protected]
> https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-users
_______________________________________________
openwrt-users mailing list
[email protected]
https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-users

Reply via email to