These repairs should have lasted from 6 p.m. to 9 p.m. Friday night.
But just as they began work, part of the hardware crashed, which was followed by a failure in another part of the Unix system—the part which contained, among other files, the data and home directories for accounts beginning with the letter ‘m.’
This second emergency was completely unrelated to the core-misc1 failure that had prompted HASCS to start repairs, according to Davis.
“It’s as if you were trying to extinguish a fire in your kitchen, when all of a sudden there was a fire in some other part of your house,” he said.
At that point, HASCS “ceased all work at once,” Davis said. “This is just not something that is supposed to happen.”
—Staff writer Eugenia B. Schraa can be reached at schraa@fas.harvard.edu.