The Supercomputing Steering Committee advises Hydra's users that, in the event of a significant hardware failure, Hydra may not resume operation, or its down time may be protracted-several days or weeks. Please plan accordingly. At the very least, you should back up valuable data on your hydra:/scratch directories on a different system. (Home directories are backed up daily.)

Hydra, our 5-year old p5 cluster, has been operating without hardware or software maintenance support for the last two years. Failure of a critical component carries the risk of forced decommissioning or protracted down time. For that reason, we advise you to plan your computational work around these two contingencies.

Posted on: 10:36 AM, February 1, 2012