Selective filtering of multi-address answers
andris at hpl.hp.com
Mon Jun 11 22:23:29 UTC 2012
On 6/11/2012 2:54 PM, Mark Andrews wrote:
> you should also be pushing for proper multi-homed server
> support in those applications that are causing you problems (read
> just about all IP applications). This is relatively easy for TCP.
I agree. Our networking group has plans to push out IPv6 capability
as one of the Next Big Projects and applications really need to have
some kind of Happy Eyeballs capability. I used your select(2) code
in a legacy application that I ported to IPv6 and it works very well.
Of course, this would also address the general problem of clients
timing out on a bad multi-address answer regardless of IPv4/v6.
More information about the bind-users