I don't have any idea where the actual problem might be. For the record, I have no DX anymore, so I can't test that myself. (As a related note, I don't have enough spare time, either.)
Well, it all boils down to plain logic thinking. #1: Find the problem, #2: solve it. #1 is the hardest, except when you tend towards perfectionism, then you can make #2 arbitrary hard, too.
As for finding the problem, you could make friends with tcpdump (and finding one for the KDX...), strace, ...
Don't fool yourself. It might be hard. You need a trained eye and brain to find some kind of programming errors. I wouldn't even consider my eyes to be fully trained yet. Such a thing takes a live of active training.
I would suggest some kind of a band-aid (AKA dirty patch) here, but I have a hard time thinking of one. Constantly restarting in, say, 1-minute intervals, is probably annoying when everything is well and is probably a too long interval for when it breaks (with a medium of 30secs to remedy). Well, so it has to be a tad bit more sophisticated.
First, find out if it's the SSH tunnel that is broken. Does the terminal window still react? If not - make use of that. Detect that situation and break the connection. You would need to script the connection on the Windows side for that. Thousands of ways to do so (OK, exaggerating a bit), but maybe start with using a command line SSH client rather than PuTTY, see if it has options to error out and quit for when the connection breaks (some of them do!). Put it in a loop for reconnection.
And so on. I just can't give a simple recipe, sorry.
|