Due to a power failure on part of the rack fimm was unavailable from 03:00 to 08:40. Most jobs failed due to one network switch being without power, some jobs may still be running on the nodes without being visible in the queue. We are looking into this and cleaning up.