Currently the way mirc handles dns pools it will pick 1 ip out of the pool and stick to it, so if it fails to connect instead of picking another ip it throttles the same ip again and again.

example

irc.network.com

127.0.0.1, 127.0.0.2, 127.0.0.3

ok lets say 127.0.0.2 server goes down and mirc has picked that ip.

The user would get the impression the network is down as the user would keep getting connection refused message instead mirc could simply move to another ip address in the pool and connect ok.

Could mirc be made to handle this differently so on every connect attempt it looks up the address and picks the ip out, perhaps this is linked to the /dns command that only outputs 1 ip hence the problem.