Murray Hill suffered a power oddity, I think a spike, on January 1st that caused a several-hour-long outage and seems to have caused miscellaneous other damage. We've had strange network problems since. Friday was the first day that things were more-or-less back to normal, at least from an external perspective.
Much of the content of cm.bell-labs.com, notably /cm, is served to the outside through a small set of machines, all of which have been replaced in the last few days. In particular, plan9.bell-labs.com now has much more RAM, so we shouldn't need to restart the wiki as often. We usually notice problems with the outside machines fairly soon (more so on weekdays), though fixing them sometimes takes a while, depending upon the problem.
