Jobs on the OSPool run on machines all over the country that cannot see your access point's files. Each job has to bring its inputs along and send its outputs back. HTCondor does that for you once you say which files to move.
1. Where to put files
| Location on the access point | Use it for | Default quota |
|---|---|---|
/home/your_username |
Scripts, submit files, inputs and outputs under 1 GB per job. Always submit jobs from here. | 50 GB |
/ospool/apNN/data/your_username |
Inputs or outputs over 1 GB, inputs shared by many jobs, and container images (.sif) |
100 GB |
/ospool/uw-shared/projects/PROJECT |
Files shared with your OSPool project group | varies |
apNN is your access point (for example ap40). Quotas are from the OSPool data overview; your login message shows the current ones, and the OSG takes requests for more.
2. Get files onto the access point
From your laptop or the HPCC, use scp or rsync (WinSCP works on Windows):
rsync -av my_inputs/ your_username@apNN.uw.osg-htc.org:~/project1/
rsync -av big_reference.tar.gz your_username@apNN.uw.osg-htc.org:/ospool/apNN/data/your_username/
For very large datasets, talk to Research Computing about Globus or about publishing the data through the Open Storage Network.
3. Small files: let HTCondor move them
List inputs in the submit file. Outputs the job writes in its working directory come back to the submit directory automatically:
transfer_input_files = data/sample_$(Process).csv, analysis.py
4. Larger files: use OSDF paths
Files under /ospool/apNN/data/your_username are reachable from jobs as osdf:///ospool/apNN/data/your_username/.... Use that path in the submit file and HTCondor fetches the file through the OSDF caches:
OSDF_LOCATION = osdf:///ospool/apNN/data/your_username
transfer_input_files = $(OSDF_LOCATION)/reference_v2.tar.gz, analysis.py
Send a large output straight to OSDF storage instead of your home directory:
transfer_output_remaps = "results.tar.gz = $(OSDF_LOCATION)/results_$(Cluster)_$(Process).tar.gz"
Caches keep old copies. Never overwrite a file under the same name. Put a version or date in the name (reference_v2.tar.gz, container_2026-10-29.sif).
5. Containers
Put Apptainer images (.sif) in your OSDF space and point the submit file at them, as in the OSPool containers guide.
6. Bring results home
Copy results from the access point back to UCR storage, for example to the HPCC or CephRDS:
rsync -av your_username@apNN.uw.osg-htc.org:~/project1/results/ ./results/
Access point storage is working space, not backed up storage. Keep your own copy of anything you need.
7. Reading public OSDF data from anywhere
The pelican client reads public OSDF data from a laptop, the HPCC or a notebook, with no account:
pelican object ls osdf:///ospool/uc-shared/public/OSG-Staff/validation/
pelican object get -r osdf:///ospool/uc-shared/public/OSG-Staff/validation/ ./validation
Next: Submitting jobs and getting results.