|
Joseph Mack NA3T wrote:
> On Thu, 29 Sep 2005, Dan Trainor wrote:
>
>> My question was, how do people provide these high availability systems
>> that I speak of?
>
>
> The solution we've got (and I'm glad to hear of any others), is to LVS
> the nameservers, and to failout a realserver that's not working.
>
> Failing out DNS using mon is covered in the HOWTO.
>
> Joe
Joe -
LVS'ing the nameservers solves the problem of redundancy to an extent.
However, it's the issue of being able to issue resolutions to clients in
a round-robin fashion, utilizing some sort of loadbalancing system.
All this aside, I think we've figured out a way to make this happen. In
our application, we're implementing our own resolver library which will
rely on the list of authorative nameservers returned by a root server
query, and use those accordingly. The main problem here was worrying
about other nameservers not respecting our TTL; this way, we can force
our TTL, and ensure that there are no nameservers in the "middle" that
can chose to ignore the TTL.
Thanks for the help
-dant
|
