1 |
Am Wed, 21 Sep 2016 21:29:13 +0200 |
2 |
schrieb Kai Krakow <hurikhan77@×××××.com>: |
3 |
|
4 |
> Am Wed, 21 Sep 2016 07:30:40 -0700 |
5 |
> schrieb Grant <emailgrant@×××××.com>: |
6 |
> |
7 |
> > [...] |
8 |
> > [...] |
9 |
> [...] |
10 |
> > > |
11 |
> > > If that device behaves badly in router mode by blocking just all |
12 |
> > > icmp traffic instead of only icmp-echo-req, this is a good idea. |
13 |
> > > You may want to bug AT&T about this problem then. It should really |
14 |
> > > not block related icmp traffic. |
15 |
> > |
16 |
> > |
17 |
> > Hi Kai, yesterday I switched my Gentoo router over to handling PPPoE |
18 |
> > and pings seem to be working properly now. The AT&T device is now |
19 |
> > functioning as a modem only and passing everything through. Today |
20 |
> > I'll find out if it helps with TCP Queuing and (supposedly) related |
21 |
> > http response slowdowns. |
22 |
> |
23 |
> You may want to set the default congestion control to fq-codel (it's |
24 |
> in the kernel) if you're using DSL links. This may help your problem a |
25 |
> little bit. It is most effective if you deploy traffic shaping at the |
26 |
> same time. There was once something like wondershaper. Trick is to get |
27 |
> the TCP queuing back inside your router (that is where you deployed |
28 |
> pppoe) as otherwise packets will queue up in the modem (dsl modems use |
29 |
> huge queues by default). This works by lowering the uplink bandwith to |
30 |
> 80-90% of measured maximum upload (the excess bandwidth is for short |
31 |
> bursts of traffic). Traffic shaping now re-orders the packets. It |
32 |
> should send ACK and small packets first. This should solve your |
33 |
> queuing problem. |
34 |
> |
35 |
> Between each step check dslreports.com for bufferbloat. I'm guessing |
36 |
> it is currently way above 1000 ms while it should stay below 20-50 ms |
37 |
> for dsl. |
38 |
> |
39 |
> The fq-codel congestion control fights against buffer bloat. But it |
40 |
> can only effectively work if you're doing traffic shaping at least on |
41 |
> your uplink (downlink may or may not be worth the effort depending on |
42 |
> your use-case). |
43 |
> |
44 |
> Additionally, you can lower the priority of icmp-echo-reply this way |
45 |
> so during icmp flooding your uplink will still work. |
46 |
> |
47 |
> This link may help you: |
48 |
> https://www.bufferbloat.net/projects/codel/wiki/Cake/ |
49 |
|
50 |
And this: |
51 |
https://github.com/tohojo/sqm-scripts |
52 |
|
53 |
-- |
54 |
Regards, |
55 |
Kai |
56 |
|
57 |
Replies to list-only preferred. |