brctl setfd virbr0 0.1 by default?
Chuck Anderson
cra at WPI.EDU
Mon Jul 21 21:40:10 UTC 2008
> https://bugzilla.redhat.com/show_bug.cgi?id=435757
> Sometime after F8, something changed where stuff attached to a bridge
> fails to connect until 15 seconds later. A manual workaround of brctl
> setfd BRIDGENAME 0.1 makes stuff work immediately.
>
> Are there any reasons why don't we do this by default for virbr0 in libvirt?
>
> Use cases:
> - Attach a different tun/tap device to the virbr0 bridge to talk to your
> virtual machine.
> - Attach ethX real device to your vribr0 bridge to talk to your virtual
> machine without NAT (like if you are doing DHCP.)
I would assume that brctl or the kernel became fully 802.1d Spanning
Tree Protocol compliant. Compliant bridges MUST delay entering the
forwarding state at link up or initialization, transitioning between
LISTENING and LEARNING states first. This is required by the
standard.
That being said, most really switches support a per-port "fast start"
mode sometimes called "portfast" or "edge" where the port is
designated to be connected directly to an end system rather than
another bridge/switch, which causes the forwarding delay to be set to
0. This is not strictly compliant to the standard, but is often used
because of the issue where DHCP clients will time out after 15
seconds.
More information about the fedora-devel-list
mailing list