Date
Projects
- NM Cluster scheduler replacement
- ALMA profiling
- Alamo Starlink
- GPU support
- NAASC scheduler replacement plan
- ngVLA Data Processing ConOps
- VLASS, assist with cube imaging plan
Discussion items
| Item | Who | Notes |
|---|---|---|
| HERA hardware | James |
|
| Slurm bug | krowe, jrobnett | Respond to SchedMD's question about suport contract. |
| HTCondor show and tell | krowe | Make a show-and-tell to Amy and the DAs. Thursday, Mar. 17 at 9am. Make a slide about how to submit jobs manually. Use James's zoom. Tell Amy and the DAs |
| VLASS cron job | krowe | Done: VLASS wants a cron job set up to copy data hourly, putting here so I don't forget. In place but a pgrep -f /lustre/aoc/projects/vlass/bin/pb-psf-storage.py would be good to keep it from running twice. Problem is cron won't let me use the -f Of course you can't use pgrep -f in a crontab entry because it will always find itself. So I wrote a wrapper script. It now checks for itself before running. |
| DMS computing plan | James | Review DMS computing plan https://sharepoint.nrao.edu/dms/_layouts/15/WopiFrame.aspx?sourcedoc=/dms/Shared%20Documents/Management/Science%20Information%20Services/DMS%20Scientific%20Computing%20Management%20Plan.docx&action=default |
| RADIAL plan | James | We need to characterize cluster computer services from the perspective of what can be supported by the remote host and what needs to be supported by NRAO |
| VLASS SE memory | krowe | Test an SE continuum imaging job on a dedicated node to see what the memory footprint should be. It's currently 50GB could it be 32GB with swap? Ask Amy for a test job. |
Action items
CAS-13342