So the dead cluster node in public tests was due to us having killed the data hard drive (every read/write results in device errors) 
my guess is it has not enjoyed hundreds of power outages per day the last few weeks.
And Murphy’s law would say it happens on the first open day of cluster testing. But hey the cluster did its job and kept running.
We are changing the drive on that node and it will be up soon for people to watch.
Greg
