I've been thinking it through a bit after seeing the pool stratum connection disconnect at every start of a new round. Seeing my miner go down for 15-30 seconds, and about 45 seconds till they are back up to speed.
I am wondering are you restarting a new stratum server each time it switches coins?
If so I wonder if it would be more efficient to have a stratum proxy (or 2, one in NA) running with a really fast reconnect rate to the pool stratum server ... and have all clients maintain connections to that proxy, seems to me that doing this may might get a faster work restart after the coin switch.
Its been bugging me in my head that I see stuff like this quite often in my miner output on the coin switch. This one happened right on the finding of a block ... did it even get found, or was that one of the 24 shares lost due to disconnect?
[2013-09-28 14:25:23] Accepted 7009652c Diff 1.12K/20 GPU 0
[2013-09-28 14:25:34] Found block for pool 0!
[2013-09-28 14:26:57] Stratum connection to pool 0 interrupted
[2013-09-28 14:26:57] Lost 24 shares due to stratum disconnect on pool 0
[2013-09-28 14:27:19] Pool 0 stratum+tcp://stratum01.hashco.ws:8888 not respond
ing!
Hmm ... as I type this it seems it is a much longer than normal disconnect
.