+Open all         -Close all

• About DCSC/SDU
  • Home
  • Board
  • News
• Resources
  • Common infrastructure
  • Cluster: Horseshoe9
• Support
  • How to Get a Grant
  • How to Get an Account
  • Docs
    • FAQ
    • Intel MKL
    • Modules
    • MAUI
    • Slurm
• Contact
  • DCSC/SDU Contact
  • DCSC Central Office
  • Other DCSC Centers

 The user homedirectories are located on mirrored RAID5 boxes for the protection
 of userdata - this homedirectory is available on all the execution hosts and 
 frontends in the cluster. However since the network does not support a great
 deal of traffic, users are required to use the scratch space "/scratch" on
 the execution hosts for temporary data.
 When a job completes, output from the computation can/must be copied to the
 users homedirectory or "/wkspace" (read on). The information about PBS
 will show some examples of this. 

 *PLEASE NOTE*
 After the completion of a job, "/scratch" will be erased on the execution
 nodes used by that job. 

 The user has access to 2 other pools of diskspace on all execution nodes and
 frontends called "/wkspace1" and "/wkspace4" meant for storage of temporary
 data. These areas have a capacity of 4,6TB and 76TB respectively.

 Current usage on the homedirectory and workspace filesystems is available on
 the Disk Usage web page.

 DCSC/SDU does not offer diskspace for long term data archival.

 *PLEASE NOTE*
 We do not backup any of the filesystems in the Horseshoe cluster. If a
 filesystem crashes, your data is lost unless you have evacuated them to
 safe place outside DCSC/SDU.