Glad IW have looked into it and the issue is now known.
I’m sorry, as I said, we do not use IPV6 currently, but as I was working away for a few days, I thought it might be gateway as issue, clearly wrong but as I do not access other systems, is hard to know full facts or run tests.
We had similar when setting up multiple networks on same server, and it was easy to fix using routes, as the system did not know which route to take, on which gateway for the new IP’s (IPV4). Once we had told the system, it worked lovely, which if it helps and you want to try as test, below is what I believe you need to use. I’m sorry if I am wrong. The first part should tell you which nic (device) is used by IPV6, assigned by RA, which you can then use this to set the IP routes with, so the system knows which gateway/route to use.
The above though, is only my thoughts, which I am sure IW have already tried/thought off, as those guys rock.
Lastly, as I think if your system is using IPV6 (well setup for IPV6 and IPV6 works in SSH), I would check your email server to make sure it is not using your IPV6 address, if it is, I would either disable it or make sure you have RDNS set on it. Qmail would use IPV6 first if available over IPV4, even if you set the Qmail to use only IPV4.
To test, send an email from your IW server and look at the header details.
I hope that helps a little
/route -A inet6 - to find all IPV6 routes/device
route -A inet6 add <ipv6network>/<prefixlength> gw <ipv6address> [dev <device>]
ip -6 route add <ipv6network>/<prefixlength> dev <device> src <ipv6network> table admin
ip -6 route add default via <ipv6network-Gateway> dev <device> table admin
ip -6 rule add from <ipv6network> table admin
ip -6 rule add to <ipv6network> table admin
ip -6 route flush cache