Planned downtimes

This page summarizes planned/on-going work within the Atlas cluster along with a few details. Usually, we will issue condor_off -peaceful at least 12h in advance, which means, no new jobs will be scheduled to that node and currently running jobs have another 12h to finish. However, after this, we usually need to kill the jobs on the machines to start with our work.

If you want to cross-check which nodes may have been affected, we use a simple scheme denoting the location of a node within the cluster room, e.g. if you see the IP address 10.10.7.18 in your log files, this refers to the node named a0718 which is a server in rack 7 in height unit 18.

date	racks/nodes	status/notes
2019-07-18/19	racks 5&6	done
2019-07-25	racks 1&2	done
2019-07-26	racks 3&4	done
2019-08-02	racks 7&8	done
2019-08-05	rack 9	done
2019-08-06	rack 10&11	done
2019-08-07	rack 12 (caution: special servers in rack 12)	done
2019-08-08	rack 17	done
2019-08-09	rack 18 (caution: special servers in rack 18)	done
2019-08-13	rack 19	done
2019-08-15	rack 20	done
2019-08-16	rack 21	done
2019-08-19	rack 22 (caution: special servers in rack 22)	done
2019-08-19	rack 23	done
2019-08-23	rack 24	done
2019-08-23	rack 27	done
2019-08-23	rack 28	done

-- SvenRenas - 24 Aug 2019