Galileo back to production – IMPORTANT: new scratch filesystem

  1. /
  2. HPC Center news
  3. /
  4. Galileo back to production...

Dear Users,
Galileo is now back to full production.
PLEASE NOTE: a new, larger, scratch filesystem is now online at the mount point /gpfs/scratch. The previous scratch areas are accessible, in READ-ONLY fashion, at the mount point /gpfs/scratch_old//, with the typically being userexternal. The new variable $SCRATCH_OLD was defined for each user to easily access the old scratch data.
 
The new user scratch areas are accessible via the usual $CINECA_SCRATCH variable, and they are empty. Users have to copy the data needed for future jobs from $SCRATCH_OLD to $CINECA_SCRATCH, and to save elsewhere or transfer to local storage resources what is not needed but cannot be discarded. The scratch_old filesystem will remain available for one month (until 16/07/2020).
 
Please note that the jobs running in $CINECA_SCRATCH and needing data from $SCRATCH_OLD will fail. For pending jobs of this type, you can hold them with “scontrol hold ” the time needed to copy the relevant data, and you can then release them with “scontrol  release “.
 
Best regards
 
HPC User Support @ CINECA