Jobs & Execution
A job is the basic unit of computing on Rescale. It packages input files, software, hardware, and a runtime, runs unattended, and saves its results back. To watch or interact with something while it runs, use a workstation instead.
New to Rescale?
Run your first job
A 20-minute walkthrough, from a blank job to downloaded results, using a real OpenFOAM case.
How a job runs
What Rescale provisions on your behalf, and why jobs behave the way they do.
Set up and submit
Create and submit a job
The five setup steps: inputs, software, hardware, automations, review.
Find and manage your jobs
The Jobs list, filters, saved views, search syntax, columns, and CSV export.
Reuse a job setup
Duplicate a job, or save a template your whole team can launch from.
Job priority
Reference: ODP, ODE, and ODR, what each means for cost and interruption risk.
While it runs
Monitoring & status
Live-tail output, check cluster health, and read the logs.
Stop a job
Graceful stop versus force stop, and what you lose with each.
Why a job is queued
The checks a job must pass before it launches, and what requeuing means.
Checkpoint long simulations
Make a multi-day run survivable, and resumable.
Afterwards
Download your results
Get files off the platform, including multi-run jobs and sets over 4 GB.
Archive and delete jobs
Stop paying for storage you no longer need, and the trash step people miss.
Troubleshooting
Diagnose failed, stuck, or killed jobs.
Working with others
Share a job
Give a colleague (or Rescale Support) access to a job and its files.
Comment on a job
Discussion attached to the run itself, with mentions and threads.
Related
- Workstations — the interactive counterpart
- Workflows & DOE — parameter sweeps and multi-run studies
- Files and storage — where job data lives
- Software catalog — solver-specific guidance