Mining server restart to resolve bug

There's a bug in the mining server where when a miner using the default Equihash (200,9) algorithm to the Equihash (125,4) algorithm, a server can lock up. The cause of the problem is that the size of the default equihash solution is 26 times larger, so a lot of data gets unnecessarily sent to the share verifier for the (125,4) algorithm.

We added a check to reject shares with the wrong solution size in this new version, so that the share verifier can't get overloaded and cause the mining server to lock up.

The restart will disconnect miners for about five minutes.