As of yesterday a new usage policy has been implemented on Tödi. This is a development system therefore it can be used as follows:
**Prime Time 8AM – 6PM : no production runs are allowed. Only jobs with a wall clock time of 1 hr can be submitted for development purpose.
**Non Prime Time 6PM – 8AM + Weekends: free for production runs.
The next regular scheduled maintenance day for CSCS systems is Wednesday July 6th.
On that day the following systems will be affected, with the expected downtimes as indicated:
* Front-end system:
Ela: 8:00 -12:00 all nodes will be alternatively rebooted
* Cray XT5:
Rosa: 07:00 – 19:00 HW and SW upgrade
* Cray XE6:
Palu: 08:00 – 18:00 SW Upgrade
* Viz and post-processing:
Eiger: 08:00 – 12:00 SW upgrade
Rigi: 08:00 – 12:00 SW upgrade
Julier: 08:00 – 14:00 SW upgrade
* File System and Storage:
/home , /apps, /project: 9:45 short intervention, downtime no more than 30 minutes
Network no maintenance scheduled.
IMPORTANT:
Be aware that beginning on Wednesday , the /scratch file system on Palu will be regularly cleaned of old data. We will apply the same cleaning policy currently on rosa. That is all files that are older than 30 days will be deleted.
In the event that cleaning files older than 30 days still leaves the file system filled to greater than 80% of its capacity, CSCS staff will take further steps to free up space, and in particular we will begin by focusing on the largest users of the /scratch file system.
Please note that scratch file systems are intended to be used for temporary storage while a job is running, and important data should be copied onto alternative storage facilities as soon as a job has completed.
Additionally, we are monitoring the usage of the “touch” command both on rosa and palu. Please note that anyone found to be using this or any other commands to prevent the cleaning policy and to keep using /scratch for long-term storage will face serious consequences and will not be allowed to run any jobs for the remainder of the allocation period.