How Tos

This section provides step-by-step instructions for specific tasks.

The torc CLI toolkit provides the simplest mechanism to build and manage workflows. It provides most functionality and this section describes it. If you need or want more control, you are welcome to use the API through OpenAPI-generated libaries or API tools like curl, Postman, and Insomnia. You can also use Arango tools to manage data directly in the database.

• Getting Started
  • ArangoDB instance
    • Start the ArangoDB Docker container
    • Create and configure a database for your workflows
  • Installation
  • CLI Fundamentals
    • General Usage
    • Database Connection
    • Workflow Key Shortcuts
    • User Confirmation
    • Output Format
    • Shell Completion
  • Workflows Quick Start
    • Create a workflow
• Managing Workflows
  • Configure a workflow
    • Configure a workflow specification
    • OpenAPI Clients
    • Startup and completion scripts
    • Compute node configuration options
  • Run a workflow
  • Check workflow status
    • TUI
    • Monitor events
    • Job status
    • Return codes
  • Cancel a workflow
  • Restart a workflow
    • Dry run
    • Restarting jobs that did not finish
    • With job dependencies
    • Without job dependencies
    • Restarting jobs with –only-uninitialized
  • Automated scheduling
  • Gracefully shutdown jobs
  • Debugging
    • Log files
    • Slurm error messages
    • Common Problems
  • Passing data between jobs
  • Plot Graphs
    • Pre-defined graphs
    • User-defined graphs
  • Backups
    • Arango tools
    • SQLite
    • Workflow Specification
  • Archive a Workflow
• Results
• Compute Node Environments
  • HPC Compute Node Environment
    • Slurm Configuration
  • HPC Multi-Node Jobs
    • One torc worker application
    • One torc worker application for each compute node
  • Custom Invocation Environments
  • Run Jobs on an Arbitrary Computer
  • Slurm Hardware Control
    • Slurm CPU management
  • Set CPU Affinity
• Compute Node Utilization
  • Resource Utilization Statistics
    • Configuration
    • Aggregated Stats
    • Time Series Stats
  • Node Packing
  • Auto-Tune Resource Requirements
  • Parallelization within a compute node
• API
  • API Access With cURL
    • Access through torc-service API
    • Access through ArangoDB HTTP API