Uppsala Multidisciplinary Center for Advanced Computational Science

Many Fysast1 and Milou nodes crashed on Saturday morning (2016-10-22) -- NOW BACK IN PRODUCTION


A total of 146 compute nodes of Fysast1 and Milou crashed on Saturday morning at about 0600 hours. It looks like a problem with the power supply. The nodes are now back in business.

If you ran a job on any of these system at that time, it may have been affected, meaning that you need to resubmit your job. You can see it with command finishedjobinfo: "jobstate" will be "NODE_FAIL".

The affected nodes were:

m[10,16,18,30,36,60,70,72,83-85,88,92-94,146,170-171,174,178-179,183,189-190] m[2-8,12,14,25-26,28-29,31-34,38,40,42-46,48-54,56,58,62-63,66,69,73-75,77-78,80,82,90,97-98,100,102,105-109,112,114,116-118,138-145,147-160,162,168,172,176,181-182,184,192,196,198,200,202,204,206] m[211,213-216,218,220,222-224,228-230,232-237,239-244,246,248]

We are sorry about any inconvenience you may experience due to this crash.

Old System News