The stratum proxy doesn't handle any error. Doesn't matter what you send, it forwards. As I told you another problem is, that you introduces a single point of failure. I don't want my whole cluster offline, because my pi proxy is down. From my network, just the first worker is able to connect to mint.bitminter.com. In all the others, the bfgminer switches to a backup pool (it says that mint.bitminter.com is dead) I have 3GB/H going to you and the others 12GB/H going somewhere else :-/
That's odd. If one can keep mining then the others should work too.
pi@raspberrypi ~ $ nc -v -w 5 mint.bitminter.com 3333
nc: connect to mint.bitminter.com port 3333 (tcp) timed out: Operation now in progress
This works fine for me, with other connections already established.
pi@raspberrypi ~ $ nc -v -w 5 mint.bitminter.com 9332
nc: connect to mint.bitminter.com port 9332 (tcp) timed out: Operation now in progress
Here the port number is wrong. getblocktemplate and getwork is on port 8332, not 9332. Alternatively Stratum is also on port 5050 and gbt/getwork on port 80.