Re: Clarification in changing the router id dynamically!!

[Acee Lindem <[e-mail]>]

kamatchi soundaram wrote:
> Hi Acee,
>
> ----- Original Message -----
> From: "Acee Lindem" <[e-mail]>
> To: <[e-mail]>
> Sent: Friday, July 04, 2003 6:25 PM
> Subject: Re: Clarification in changing the router id dynamically!!
>
>
>
>>kamatchi soundaram wrote:
>>
>>>Hi,
>>>
>>>  If the router id of the router R1 is changed on the fly (dynamically),
>>>what factor triggers the adjacent router R2 connected to R1 via point to
>>>point link, to generate the new Router LSA???
>>>
>>> From the RFC i could infer that R1 will flush all his self originated
>>
> lsas
>
>>>when it's router lsa has been changed.
>>>  But i couldn't find, how the R2 triggered to generate the new router
>>
> lsa
>
>>>with the changed router id as is neighbor id in its router lsa??
>>
>>Kamatchi,
>>
>>Actually, since neighbors are identified by router ID on P2P links (as per
>>RFC 2338 section 8.2), R2 shouldn't recognize R1 with the new router ID
>
> and
>
>>should bring down the old adjacency with R1 and form a new one. A more
>>interesting question is what would happen on a broadcast or NBMA network
>>(where neighbors are identified by source address).
>>This case isn't explicitly handled by RFC 2328 and there is some variance
>>between implementations. At a minimum, the DR must recognize the change
>>in router ID and originate a new network LSA. I found that one popular
>>implementation required the router with the router ID change to bring its
>>adjacencies down (by sending a hello w/o any neighbor IDs) in order to
>>interoperate.

Hi Kamatchi,

>
>
> Does it mean that the R1 (router whose Id has got changed), should reset
> it's NSM with all its neighbor to down/init state and then send hello
> without any neighbor ids??
>
>   Since, R1 need to flush all it's self originated lsas before to bring down
> the adjacency and restart again, there needs some ample amount of time to
> flood and flush the lsas in the routing domain (atleast with the neighbor).
> So, i cannot send the hello packet (without any neighbors) immedietly after
> the sending out the max age lsas (to flush the self originated lsas).

I purge self-originated LSAs before sending the empty hello. I do both
before the router ID change. I do not add the self-originated LSAs to
any retransmission lists and I do not wait for acknowledgements. Some
implementations just let their old LSAs time out and without any
strong complaints.

>
>  Because if i do that, then there is a chance that before the self
> originated lsas are completely flushed out or sent from R1, the NSM may
> reverted to down state. Which will basically prevent the complete flushing
> of self oringinated lsas . . correct??

Only if your Link State Update packet is dropped.


>
>  My question is how can be the time between the flushing the lsas and
> hello's sent without the neighbor ids determined???
>
>  Since, if there is no significant amount left between the lsa flushing and
> new hello without neighbor id, then the entire system will be in bad
> condition.
>
> Thanks,
> Kamatchi.
>
>
>>Thanks,
>>Acee
>>
>>
>>>Thanks in advance,
>>>Kamatchi.
>>>
>>
>>
>>--
>>Acee
>>
>
>


--
Acee