Flux CORAL2

This project contains plugins and other add-ons required for Flux to run as the primary resource manager on HPE Cray EX systems, specifically the CORAL-2 systems at Lawrence Livermore National Laboratory:

Name	Node Count	APU
tuolumne	1,152	MI-300A
rzadams	128	MI-300A
elcap	11,136	MI-300A

Plus these early access test systems:

Name	Node Count	APU
tioga	32	MI-250X
rzvernal	38	MI-250X
tenaya	24	MI-250X

Documentation is currently under construction.

Table of Contents

• Manual Pages
  • Section 1 - Flux CORAL2 Commands
  • Section 5 - Flux CORAL2 Configuration File
• Flux CORAL-2 Administration
  • Background
  • Installing Software Packages
  • Overlay Network Configuration
  • Slingshot Configuration
  • Enabling Cray MPI
  • Rabbits
• Slingshot Interconnect
  • VNI Tagging
  • Traffic Classes
  • NIC Resource Management
  • Hardware Collective Offload
  • Implementation
• Rabbit Near-Node Storage
  • Background
  • Components
  • Workflow State Machine
  • RPC Interface
  • Resource Scheduling Integration
  • Error Handling
  • Implementation
• Flux with Rabbits
  • How to Allocate Rabbit Storage
  • Preset DW Directives
  • Custom DW Directives
  • Enabling Rabbit Fault Tolerance
  • Fetching Rabbit Information
  • Additional Attributes of Rabbit Jobs
• Glossary
• References
• flux_k8s package
  • Submodules