Mandi! Andy Furniss
In chel di` si favelave...
> I have no experience of WiMAX, but it may be different enough from
> "normal" links to cause problems.
> If it's rated 3M down then the 3mbit ceils and rate on parent will
> be too high.
As every wireless link, proven to be a bit unreilable... and it is not
a PtP link, so all ''bad'' effect of a shared link added.
Anyway, better that wifi/5GHz...
Now i've set in bandwidth to 2816 Kbit/s. Seems effectively better...
> >ifconfig ifb2 mtu 1476 txqueuelen 600 up
> I wouldn't bother setting mtu here (though the test I just did
> looked like it didn't do anything anyway). The txqueuelen seems
> long, but as you add sfq you'll get the default for that (128).
But interface seems correctly defined:
tank:~# ifconfig ifb2
ifb2 Link encap:Ethernet HWaddr ae:39:fc:54:66:5f
inet6 addr: fe80::ac39:fcff:fe54:665f/64 Scope:Link
UP BROADCAST RUNNING NOARP MTU:1476 Metric:1
RX packets:7774590 errors:0 dropped:3 overruns:0 frame:0
TX packets:7774587 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:563
RX bytes:3971715041 (3.6 GiB) TX bytes:3971714831 (3.6 GiB)
for the queue lenght.. you're right, see later.
for the tx buffer, you mean that ''doesn't bother'', or i've to set
accordingly the sfq buffer with the 'limit' paremeter?
But i don't define a sfq for the parent, i've to setup the limit
parameter for every children?
> >/sbin/tc class add dev ifb2 parent 1: classid 1:1 htb rate 3000kbit ceil 3000kbit burst 16k cburst 8k mtu 1476
> I also wouldn't set mtu here it's not the same meaning as doing it
> on a real if - IIRC it also needs 14 adding for traffic on/from eth)
This is true also for the egress shaping, for eth2? Eg, setting mtu on the parent
class for eth2 mean nothing?
> >/sbin/tc qdisc del dev ifb2 ingress
> >/sbin/tc qdisc add dev ifb2 handle ffff: ingress
> >/sbin/tc filter add dev ifb2 parent ffff: protocol ip prio 50 u32 match ip src 0.0.0.0/0 police rate 3000kbit burst 16k drop flowid :1
> ingress doesn't work on ifb so I would get rid of this.
But again seems that works... ops, no, sorry, you're right.
tank:~# tc -s -d qdisc show dev ifb2 ingress
[...]
qdisc ingress ffff: parent ffff:fff1 ----------------
Sent 0 bytes 0 pkt (dropped 0, overlimits 0 requeues 0)
rate 0bit 0pps backlog 0b 0p requeues 0
I've reduced burst and cburst value, and seems i've resolved many trouble;
Originally i setup these parameters statically (usually, 8/4 or 16/8),
but some years ago i've started to compute them as a fraction of the
bandwidth.
Probably in the meantime connection passed from Kbit/s to Mbit/s, and
sometimes that was broken but not so evident became really broken (and
evident).
What are the best strategy to compute burst/cburst and buffers, if
there's one?
Again, thanks.
--
dott. Marco Gaiarin GNUPG Key ID: 240A3D66
Associazione ``La Nostra Famiglia'' http://www.sv.lnf.it/
Polo FVG - Via della Bontà, 7 - 33078 - San Vito al Tagliamento (PN)
marco.gaiarin(at)sv.lnf.it tel +39-0434-842711 fax +39-0434-842797
Dona il 5 PER MILLE a LA NOSTRA FAMIGLIA!
http://www.lanostrafamiglia.it/chi_siamo/5xmille.php
(cf 00307430132, categoria ONLUS oppure RICERCA SANITARIA)
--
To unsubscribe from this list: send the line "unsubscribe lartc" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
