There'll be a pool restart at 12:00 UTC in half an hour.
It will mean a failover and an immediate failback for ALL miners.This is an update of the mining pool code to include segwit transactions in our work (since they clearly have been working on the network for a while now) and a fix for the deadlock long ago back in Feb

I've also added a log dump of the block information so I can find out what these valid 'inconclusive' blocks we are getting are.
Unfortunately the update failed dismally

I switched back to the old code again and all is OK again.
Total outage was 2 minutes.
I terminated the pool at 21:59:56, but then had to switch back to the old code, all nodes had reconnected by 22:01:51 so miners would have failed back from then.
More work to do.