[Bridge] Bridge ports needlessly going through learning stage when STP is disabled

[Bridge] Bridge ports needlessly going through learning stage when STP is disabled

[rahul bhardwaj]

Hi,

I've noticed that in Linux 2.6.15, that even when I have STP disabled on a bridge (which I think is the default), when you add a new interface to the bridge, it still goes through the learning stage and shows STP state display such as TOPOLOGY_CHANGE TOPOLOGY_CHANGE_DETECTED". For this time interval box is unreachable from outside. I have verified this with the help of ping. I found that ping reply stops for some time around 15 to 30 secs, if you add another interface to the bridge. Which is causing some problems for the box.

Is this the right behavior ? The 15 to 30 second delay before the
port moves to Forwarding state can be frustrating, as you don't expect
it, and for that period of time, start to think there is something
else wrong in the network's configuration. Is there any place in the code where I do some change to resolve this issue.


Thanks and regards,

Rahul B




 


      Add more friends to your messenger and enjoy! Go to http://messenger.yahoo.com/invite/

Re: [Bridge] Bridge ports needlessly going through learning stage when STP is disabled

[Francesco Dolcini]

rahul bhardwaj wrote:
> Is this the right behavior ? The 15 to 30 second delay before the 
> port moves to Forwarding state can be frustrating, as you don't
> expect it, and for that period of time, start to think there is
> something else wrong in the network's configuration. Is there any
> place in the code where I do some change to resolve this issue.

At the moment the normal way to handle this is "brctl setfd $BR_DEV 0"

-Francesco

Re: [Bridge] Bridge ports needlessly going through learning stage when STP is disabled

[rahul bhardwaj]

[-- Attachment #1: Type: text/plain, Size: 1288 bytes --]

But STP is disabled on the bridge interface, then also "setfd to zero" it will work? Or should I enable the STP first?

With Regards 

Rahul Bhardwaj 

 

--- On Mon, 17/11/08, Francesco Dolcini <fdolcini@sysnetsistemi.it> wrote:
From: Francesco Dolcini <fdolcini@sysnetsistemi.it>
Subject: Re: [Bridge] Bridge ports needlessly going through learning stage when STP is disabled
To: "rahul bhardwaj" <rahul_bhardwaj36in@yahoo.co.in>
Cc: "Stephen Hemminger" <shemminger@vyatta.com>, bridge@lists.linux-foundation.org
Date: Monday, 17 November, 2008, 1:56 AM

rahul bhardwaj wrote:
> Is this the right behavior ? The 15 to 30 second delay before the 
> port moves to Forwarding state can be frustrating, as you don't
> expect it, and for that period of time, start to think there is
> something else wrong in the network's configuration. Is there any
> place in the code where I do some change to resolve this issue.

At the moment the normal way to handle this is "brctl setfd $BR_DEV
0"

-Francesco
_______________________________________________
Bridge mailing list
Bridge@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/bridge



      Add more friends to your messenger and enjoy! Go to http://messenger.yahoo.com/invite/

[-- Attachment #2: Type: text/html, Size: 1777 bytes --]

[Bridge] Bridge is not forwarding packet on correct vlan port <Help Needed>

[rahul bhardwaj]

[-- Attachment #1: Type: text/plain, Size: 1759 bytes --]

Hi
All,

I
am facing one very strange problem with bridge in our wireless Access Point.
Let me explain you my setup first. I have two Linux bridge br114 and br113 in
my AP. Both the bridge have two vlan interfaces under it, one Ethernet vlan
iface (eth1.xx) and another one is radio vlan interface (ath0.xx). So now both
the bridge will look like this

 

br113
 = > eth1.113 + eth0.113

br114
 = > eth1.114 + eth0.114

 

I
have connected two wireless stations with this AP. Two stations (STA113 and STA114)
are connected to br113. Now I am pinging to their respective DHCP server from
all the three stations. This is all about setup. 

 

After
some time, ping from STA113 and STA114 stops for their respective DHCP server.
I have checked the ARP table and found that it has a valid entry so I think the
problem lies in bridge code. I think bridge is only dropping the packet and not
forwarding it to proper interface. At this time I have put some message in
br_forward.c and found that it is putting the packet on same interface, I mean eth0.113
to eth0.113, due to this should_deliver is falling and packet is getting
dropped. 

 

Can
somebody explain me something about this behavior? I mean where I should look into
code to solve this issue. I am expecting some help on this because I am stuck
on this. One more very important thing, it happens when I am having two
bridges, things works fine with single bridge. I mean it happen only when I
have both the bridge br113 and br114. If I have only br113 then things are
working fine. 

 

Thanks
and Regards,

 

Rahul
Bhardwaj

 


      Love Cricket? Check out live scores, photos, video highlights and more. Click here http://cricket.yahoo.com

[-- Attachment #2: Type: text/html, Size: 3679 bytes --]

Re: [Bridge] Bridge ports needlessly going through learning stage when STP is disabled

[rahul bhardwaj]

[-- Attachment #1: Type: text/plain, Size: 1363 bytes --]

Thanks a lot Francesco for your help but still it is not working. it is still taking that much time in coming in forwarding state. Do I need to change somewhere in the code to achieve my goals?

With Regards 

Rahul Bhardwaj 

 

--- On Mon, 17/11/08, Francesco Dolcini <fdolcini@sysnetsistemi.it> wrote:
From: Francesco Dolcini <fdolcini@sysnetsistemi.it>
Subject: Re: [Bridge] Bridge ports needlessly going through learning stage when STP is disabled
To: "rahul bhardwaj" <rahul_bhardwaj36in@yahoo.co.in>
Cc: "Stephen Hemminger" <shemminger@vyatta.com>, bridge@lists.linux-foundation.org
Date: Monday, 17 November, 2008, 1:56 AM

rahul bhardwaj wrote:
> Is this the right behavior ? The 15 to 30 second delay before the 
> port moves to Forwarding state can be frustrating, as you don't
> expect it, and for that period of time, start to think there is
> something else wrong in the network's configuration. Is there any
> place in the code where I do some change to resolve this issue.

At the moment the normal way to handle this is "brctl setfd $BR_DEV
0"

-Francesco
_______________________________________________
Bridge mailing list
Bridge@lists.linux-foundation.org
https://lists.linux-foundation.org/mailman/listinfo/bridge



      Add more friends to your messenger and enjoy! Go to http://messenger.yahoo.com/invite/

[-- Attachment #2: Type: text/html, Size: 1860 bytes --]

Re: [Bridge] Bridge ports needlessly going through learning stage when STP is disabled

[Uli Luckas]

Am Monday, 17. November 2008 schrieb Francesco Dolcini:
> rahul bhardwaj wrote:
> > Is this the right behavior ? The 15 to 30 second delay before the
> > port moves to Forwarding state can be frustrating, as you don't
> > expect it, and for that period of time, start to think there is
> > something else wrong in the network's configuration. Is there any
> > place in the code where I do some change to resolve this issue.
>
> At the moment the normal way to handle this is "brctl setfd $BR_DEV 0"
>
Except that setting fd to zero might get you flooded. See thread "delay in 
bridge learning when forward delay is 0" here on the list. [1] 

regards,
Uli


[1] http://archives.free.net.ph/message/20081013.181016.4cbfa529.de.html

-- 
> Ich habe Zone Alarm in Hintergrund mitlaufen, das ist fuer mich
> Pflicht.
Ich habe eine Schüssel Weihwasser neben der Tastatur stehen. Das hilft
genauso.      (Jürgen Schmadlak)