On 01-Apr-13 05:18, Jason Healy wrote:
> On Mar 31, 2013, at 9:27 PM, Kurt Hillig <[log in to unmask]> wrote:
>> We use anycast for several other services on campus now - for example RADIUS - and it's saved our bacon more than once; I highly recommend that people take a good look at it, it's not that complex and it makes both client configuration and server maintenance much easier.
> As someone who's looking into it (and who hasn't found a great HOWTO on it), I have a question about how people have structured it.
> Do you set up the anycast manually (static routes to hard-coded next hops), or with something like OSPF? Our campus is small, and we have no topological redundancy, so we could just attach everything to the core and rely on L3-next-hop to setup small anycast groups.
> For those of you that are using OSPF or other route injection from the nodes, are you doing any service health checks before advertising the route, or just a basic "if the machine is up then it's part of the anycast group" setup? For example, not advertising a route until the RADIUS service is up, initialized, and running?
We use anycast for our DNS, with 8 recursive resolvers, spread out into
6 different locations co-located with our campus routers across campus.
We also have 5 authoritative servers similarly spread out.
We are using OSPF, via Quagga on each name server. We have some basic
health checking in place such that each node will take itself out of
routing if it is very unhealthy.
> Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/.
Participation and subscription information for this EDUCAUSE Constituent Group discussion list can be found at http://www.educause.edu/groups/.