General Documentation
- Welcome FAQ
- Secure Shell SSH
- Available Software
- Installing Software
- Guidelines and Policies
- Glossary
- Grant Support
- Sharing Data
- Containers & Singularity
- UserGroup Presentations
- Jupyter Notebook Usage
LSF Primer
Lilac Cluster Guide
Juno Cluster Guide
Cloud Resources
Backup Policy on server/node local drives
File lists
Page History
...
- GPFS and OS reserve ~12GB of RAM per host on Juno CentOS7 hosts. Usable memory for LSF jobs per node type: jx* have 240GB and ju* have 250GB
- Memory (GB) is per slot/task on Juno. Memory(GB) is per job on Luna.
- All jobs should have -W (Walltime) on Juno. LSF enforces it. Please, don't use -We on Juno.
- To check job information for DONE or EXIT jobs use "bhist -l JobID" or "bhist -n 0 -l JobID". Also "bacct" can be used. " bjobs -l JobID" should be used only for RUNNING or PEND jobs.
- There is no /swap on CentOS7 nodes. Memory is enforced by cgroups on hosts and job won't be able to use /swap. Job will be terminated on kernel level if job memory usage exceeds requested amount on LSF.
- There is no "iounits" resource on Juno.
...