On October 19th and 20th, both Triolith and Gamma will be unavailable.
During this stop NSC will perform maintenance and testing, e.g
If you have any questions regarding this, please contact NSC Support.
All login nodes (including ThinLinc) and compute nodes will be unavailable during this time, and will be restarted during the stop.
Queued jobs will only start if the timelimit requested (i.e "sbatch -t DD-HH:MM:SS") is less than the time remaining before the start of the downtime.
If you need to run jobs in the week before the downtime, you might need to shorten your jobs.
Jobs that are in the queue before the downtime will remain in the queue, and will be started normally after the downtime.
If your project has a low priority due to having run a lot recently, you might find it much easier than usual to get short jobs started in the days immediately before downtime as they don't have to compete with longer jobs. However, please note that core hours used for such jobs will still be counted, so your priority will be even less after the downtime if you take advantage of this.
If your project has a need to run wide and short jobs (i.e jobs using many nodes for a short period of time) for e.g scaling tests, this is a very good opportunity to do that. Please contact support at nsc.liu.se and describe your needs (e.g number of jobs and their sizes) and we'll get back to you with more information. Please note that such jobs are not "free", the core hour cost to your project is the same as always, but this is a possibility to run such jobs without inconveniencing other users (see https://www.nsc.liu.se/support/batch-jobs/triolith/#flatjobs)