06-15-2016 08:00 AM
What happens with anycast using OSPF? The manual says that starting and stopping the DNS service also starts and stops the OSPF daemon.
What happens if the DNS services hangs? Does the OSPF daemon shutdown, stay on or hang?
06-17-2016 03:18 PM
That's a good question and I'd be interested in the answer too.
However I haven't seen BIND hang in a very long time, if it gets upset it normally just stops, then the watchdog would start it again. Doesn't mean to say it can't happen though.
PCN (UK) Ltd
All opinions expressed are my own and not representative of PCN Inc./PCN (UK) Ltd. E&OE
07-13-2016 08:52 AM
We also had this consideration when deciding to use Anycast. What would happen in the event of some subtle failure of DNS that was too subtle to make the OSPF adjacency drop? In the event we went with the solution anyway. That was 4 years ago and have not experienced any issues with hanging processes like that. We took the view that DNS using Anycast, even with that small unknown risk, was a lot better and more flexible for us than the traditional method of primary/secondary DNS IP addresses being tied to specific members.
07-29-2016 08:41 AM
I agree with the replies. Infoblox's OSPF to named link catches most failures. We have been running OSPF anycast for more than 5 years and their have been very few problems. It over all has been a huge gain.
There have been a few bugs in the underling *nix OSPF service that Infoblox is running, 2 that I remember. One was very early on where the OSPF service would just stop and the route was then removed. So as long as you detected the failiure, before there were multiple failures and you ran out of capacity it was not an issue. The most recient was with CISCO ACI and detecting a redundant leaf swich failure. We found this in our lab testing and it was addressed quickly.