Bandwidth shaping with CBQ

From: Graeme Hinchliffe (email_at_NOSPAM-linkovitch.me.uk)
Date: 05/14/04 

Date: Fri, 14 May 2004 17:51:57 +0100

Hiya
        I am trying to setup some bandwidth shaping using the kernel CBQ modules.
 I am using the cbq.init script to configure tc for me.

        All seems well until I try to implement larger rates. I have a box with
3 interfaces, two of which are setup as a bridge, the third connects
directly to this bridge. The CBQ tree is applied to the ports on the
bridge.

        Now these ports are set to 100Mbs, and I am trying to rate limit to
60Mbps per network, so that no more than 60Mbps can be recieved from any
of the given subnets. When I apply the rule the data flow is limited to
30Mbps (approx) regardless of the rate I specify, unless I specify one the
same as the interface, in which case a full 100Mbps is transmitted.

        I have adjusted the avpkt to be 1500 (initially it was 1000 and the
throughput was 10Mbps).

The box itself has plenty of umph (dual 2.6 GHz Xeon 2GB RAM) so I can't
see that being a problem, and CPU never goes about 1%

Is there some kind of internal limit with CBQ QoS stuff?

any help resolving this much appreciated

Graeme

Relevant Pages

  • Re: distributed bridging [script]
    ... >> I'm trying to create a configuration that would bridge interfaces on ... Each host would have at least two interfaces, ... here's a completely untested script. ...
    (freebsd-net)
  • Re: bridge with access on both interfaces
    ... >>'inside' and 'outside' IPs separately on each bridge interface, ... Some tcpdumps on both in and outside interfaces suggest an ARP ... > If you want to use IP while bridging, ... and net4801 is that the 3 ethernet controllers now share one interrupt. ...
    (freebsd-net)
  • [BUG][debian-2.6.20-1-686] bridging + vlans + "vconfig rem" == stuck kernel
    ... I've managed to fairly reliably trigger a deadlock in some portion of the linux networking code on my Debian test box. ... I'm pretty sure that it's a race condition of some sort as it doesn't trigger if I ifdown the interfaces one by one, but if I run "ifdown -a" then it triggers halfway through reliably. ... Just for a bit of background on why this is so complex: When I get this networking problem sorted out I'm going to set up heartbeat and a dummy "world1" interface with a shared MAC which is added to the "world" bridge when the current system is the DHCP-client master. ...
    (Linux-Kernel)
  • Re: IPS-Builder
    ... I' d like to share with all, this script made by me based on root0 ... You need 3 network interfaces (One for management and two for the bridge) ... -This have a management interface. ...
    (Focus-IDS)
  • Re: bridge issues with pf rules on OpenBSD/Sparc
    ... > assigned so that clients on the bridged interfaces can access the ... > time, regardless of bridge state. ... > rule pass in on le0 tag t_lan ... > rule pass in on le2 tag t_wap ...
    (comp.unix.bsd.openbsd.misc)