[cisco-bba] LNS redundancy

[Oliver Eyre]

Hmm that’s odd.

It was a while ago but when I was looking into this matter, although the
virtual-access interface went down the dialer never did.

Anyway glad it worked for you.

Oliver

-----Original Message-----
From: giesen at snickers.org [mailto:giesen at snickers.org] 
Sent: Friday, 25 March 2011 5:32 AM
To: Oliver Eyre; Gary Giesen
Cc: cisco-bba at puck.nether.net
Subject: Re: [cisco-bba] LNS redundancy

My way will work to accomplish that too (and easier to remember). If the
Dialer is down (ie not connected), IOS will withdraw the route (same with
any other interface that is down).

GG
-----Original Message-----
From: "Oliver Eyre" <oliver.eyre at cirruscomms.com.au>
Date: Thu, 24 Mar 2011 10:31:05 
To: 'Gary T. Giesen'<giesen at snickers.org>
Cc: <cisco-bba at puck.nether.net>
Subject: RE: [cisco-bba] LNS redundancy

You can also use "ppp ipcp route default" on the dialer interface if you
need the default route to go away if the ppp link ever falls over.

Useful if you have redundant connections...


Oliver

-----Original Message-----
From: cisco-bba-bounces at puck.nether.net
[mailto:cisco-bba-bounces at puck.nether.net] On Behalf Of Gary T. Giesen
Sent: Thursday, 24 March 2011 4:23 AM
To: Hitesh Vinzoda
Cc: cisco-bba at puck.nether.net; Arie Vayner
Subject: Re: [cisco-bba] LNS redundancy

Your PPPoE clients shouldn't have a default route pointed at a
specific IP, but rather an Interface. On Cisco IOS, for example, it
would be "ip route 0.0.0.0 0.0.0.0 Dialer1". That way, no matter what
IP is on the other end of the tunnel, it will send the traffic down
it.

This works for any technology that is point-to-point. Don't try it on
Ethernet or you will have some fun (especially if you have proxy arp
enabled).

GG

On Wed, Mar 23, 2011 at 10:28 AM, Hitesh Vinzoda
<vinzoda.hitesh at gmail.com> wrote:
> Hi Arie,
> But the problem is both the LNS would have separate loopback ip addresses
> configured and under virtual-template we will have ip unnumbered loopback
> e.g.
> LNS-1 20.20.20.1/24 loopback and Virtual template
> LNS-2 20.20.20.2/24 loopback and Virtual template
> Now say a pppoe client has configured default route towards 20.20.20.1
> before failover and after failover it is now connected to LNS-2 whose ip
> address is 20.20.20.2, in this case pppoe client would not be able to
route
> out to the internet as the gateway ip address is unreachable.
> Another problem is connecting users whom we are going to assing /29 from
> separate pool than the virtual template1 (20.20.20.1 or .2) configured say
> 20.20.50.0/29 etc. than how we can accommodate such users on LNS...
> Please provide your inputs
> Regards
> Hitesh
>
> On Wed, Mar 23, 2011 at 6:05 PM, Arie Vayner <ariev at vayner.net> wrote:
>>
>> For PPPoE, you can make sessions hit both BRAS devices at the same time.
>> The PADI is a broadcast, so would reach all of them.
>> Each one will send a reply (PADO) in unicast, but only 1 will be the 1st
>> one, and will be used.
>> Statistically, the load will split because of CPU and load on the
devices.
>>
>> Alternatively, there is a way to introduce a delay on a specific BRAS for
>> PADO, so it will be less preferred.
>>
>> Arie
>>
>> On Wed, Mar 23, 2011 at 11:43 AM, Hitesh Vinzoda
>> <vinzoda.hitesh at gmail.com> wrote:
>>>
>>> Hi Arie,
>>> Could you please tell me how the first hop redundancy for pppoe clients
>>> would be achieved?
>>> Regards
>>> Hitesh
>>> On Wed, Mar 23, 2011 at 4:33 PM, Arie Vayner <ariev at vayner.net> wrote:
>>>>
>>>> Just ask the Telco to open L2TP tunnels to both LNSs. They can
configure
>>>> them as active/active (split the users on both LNSs) or active/passive
>>>> (depends on the LAC vendor...)
>>>>
>>>> Arie
>>>>
>>>> On Wed, Mar 23, 2011 at 9:26 AM, Hitesh Vinzoda
>>>> <vinzoda.hitesh at gmail.com> wrote:
>>>>>
>>>>> Hi,
>>>>> I m trying to achieve redundancy for gateway of pppoe adsl clients on
>>>>> LNS. My scenario is as below.
>>>>> > Telco LAC sends the PPPOE session to our LNS using L2TP to LNS. It
>>>>> > sends it to Primary LNS at all the time till the Primary LNS is
alive.
>>>>> > If primary LNS fails than Telco LAC creates tunnel to secondary
>>>>> > tunnel. All connections are teared down and sent to secondary LNS.
>>>>> > Since we know that gateway or default route configured on ADSL PPPoE
>>>>> > clients needs to unique as these are the attributes sent out by
Radius or
>>>>> > DHCP and configured manually
>>>>> > So in case if primary LNS fails than the IP address on virtual
>>>>> > template on LNS should be available to PPPOE clients. I tried
different ways
>>>>> > to achieve this
>>>>> 1. HSRP on virtual-template interface is not supported.
>>>>> 2. If i use internal physical interface on both LNS and run HSRP,
while
>>>>> having ip unnumbered fa0/0 under virtual-template configuration. The
virtual
>>>>> template takes the physical ip address instead of virtual ip.
>>>>> Lastly i thought of having anycast to have resue the situation.
>>>>> > Configured loopback interfaces on both the LNS with the same ip
>>>>> > address 20.20.20.1 and ip unnumbered command under virtual-template
>>>>> > interface. This works like charm.
>>>>> I m just wondering that is my solution correct or justifies the ISP
LNS
>>>>> scenario? or what are other methods to achieve redundany in LNS for
gateway
>>>>> ip address of pppoe.?
>>>>> Thanks in advance
>>>>> Hitesh
>>>>>_______________________________________________
>>>>> cisco-bba mailing list
>>>>> cisco-bba at puck.nether.net
>>>>> https://puck.nether.net/mailman/listinfo/cisco-bba
>>>>
>>>
>>
>
>
>_______________________________________________
> cisco-bba mailing list
> cisco-bba at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-bba
>

_______________________________________________
cisco-bba mailing list
cisco-bba at puck.nether.net
https://puck.nether.net/mailman/listinfo/cisco-bba