TTL question

[Kevin Darcy]

Barry Margolin wrote:

>In article <dhk5hs$fab$1 at sf1.isc.org>,
> Kevin Darcy <kcd at daimlerchrysler.com> wrote:
>
>  
>
>>mahermick at gmail.com wrote:
>>
>>    
>>
>>>Hi Folks,
>>>
>>>I'm trying to understand the implications of not lowering the TTL for a
>>>record prior to it being changed.
>>>
>>>Taking a hypothetical situation - I have an Primary MX record with TTL
>>>of 3600 (1 hour) and no Secondarys (for arguements sake).
>>>
>>>If I changed the MX record would the maximum time my mail would be
>>>unavailable be one hour or is it possible that it could take longer?
>>>
>>>      
>>>
>>Correct, but you are *required* to have at least 2 nameservers for a 
>>zone. So there's always going to some propagation time between them 
>>unless you make the change on all of them simultaneously (in BIND terms, 
>>that means they would all have to be configured as "type master").
>>    
>>
>
>If you use NOTIFY, the propagation time around a minute at most, 
>effectively negligible compared to the 1-hour TTL.
>
Indeed. The main point I was trying to make is that it's not *just* the 
TTL that matters when predicting when a given DNS change will become 
"visible everywhere".

                                                                         
                                       - Kevin