Hi!
> This bug doesn't seem to be fixed.
>
> # brctl setfd xenbr0 24
> # brctl showstp xenbr0|grep forward.delay|head -1
> forward delay 30.00 bridge forward delay
> 30.00
This is weird, what I fount that was happening was on kernel 3.11 and after
it, is this your case?
In that case you needed to set up the fd after setting stp, otherwise the
scripts did set fd but then when stp was set the fd was overwritten. That's
what I did on the scripts.
What can be happening here is what happened to me and took me a long time to
realize when testing this, if you have two or more STP talking bridges and
you are doing the changes on a node that is not the root of the STP, as the
root will impose the FD to the others.
The other thing I can think of is that setting the fd really quick after the
stp might still not work. To debug this please take a look at your
/etc/network/if-pre-up.d/bridge, which should look like this on -7:
if [ "$IF_BRIDGE_STP" ]
then
brctl stp $IFACE $IF_BRIDGE_STP
fi
if [ "$IF_BRIDGE_FD" ]
then
brctl setfd $IFACE $IF_BRIDGE_FD
fi
try to see how the fd is before and after setting it, if it is set but then
reverts back to 30 try to put a sleep in between and see if it makes any
difference.
Also take a look at your hello time, I don't remember what the changes done
on 3.11 RCs looked like, but they do compute things to ensure the fd makes
sense, for example, if you have a big hello time maybe your fd is forced to
30.
This is how my bridge looks like with -7:
root port 0 path cost 0
max age 20.00 bridge max age 20.00
hello time 2.00 bridge hello time 2.00
forward delay 2.00 bridge forward delay 2.00
ageing time 300.00
And that comes from this setup:
iface br0 inet static
...
bridge_ports all
bridge_fd 2
bridge_stp on
If none of this works... I need more info on your setup as I cannot
reproduce this here.
Regards.
--
Manty/BestiaTester -> http://manty.net
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
