Dear cluster users,
due to a technical problem the management node of the Horus cluster needs to be restarted. This has the following consequences for you:
1. All jobs running at the time of the restart will be killed. That will not happen immediately. We will choose the time of the restart such that as few jobs as possible will be killed (most likely Wednesday afternoon). Unfortunately, a few jobs have very long running times. We will contact the affected users tomorrow. If you are not contacted by us, your jobs are not affected.
2. All jobs that have been queued but have not started (status ‘PD’) will run after the restart. You do not have to do anything here.
3. Queuing new jobs is currently only possible in the Short Queue. We will notify you when queuing jobs will be possible again.
The Horus cluster will most likely be fully available again on Thursday. We will inform you as soon as possible.
Best regards
Your HPC Support Team

Aktualisiert um 17:56 am 21. October 2019 von Gerd Pokorra.