n0005 died o/n. This is the second time, and, again, it did came back OK. Memory ?

Lost internet connectivity from cluster room. Nikos Grigoriadis fixed it and identified the problem as being a loopback. It was probably the second 10/100 hub that was known to be problematic. Took it off.