Workflows for Day 2 Operations

View as Markdown

NVIDIA Config Manager Temporal provides comprehensive network infrastructure automation workflows. Each workflow is designed for specific operational tasks and includes detailed input/output specifications.

These workflows are not part of initial setup, which the ZTP process handles. They are designed for Day 2 operations, after devices are bootstrapped using ZTP.

The Redfish, BMC, InfiniBand, and VPC workflows are currently experimental and have not completed QA. Use them only in pilot or test environments.

Available Workflows

Network Validation

WorkflowDescriptionPurpose
Site Cable ValidationValidate network connections across an entire sitePre-deployment validation, troubleshooting connectivity issues
Device Cable ValidationValidate network connections for a specific deviceDevice-specific connectivity testing
Infiniband Cable ValidationSpecialized validation for InfiniBand networksHPC cluster validation, GPU interconnect verification
Hardware ValidationCheck device health and component statusPreventive maintenance, hardware lifecycle management

Configuration Management

WorkflowDescriptionPurpose
Configuration BackupBackup device configurationsConfiguration preservation, disaster recovery
Configuration DeployDeploy configurations to network devicesConfiguration updates, standardization
Multi-DeployDeploy configurations to multiple devicesBulk configuration deployment
Device Password RotationRotate user passwords on network devicesSecurity maintenance, credential lifecycle management
Site Password RotationRotate user passwords across all devices at a siteBulk security maintenance, site-wide credential rotation

Infrastructure Operations

WorkflowDescriptionPurpose
Switch OS UpgradeUpgrade network switch operating systemsSoftware lifecycle management
Infiniband MLNX OS UpgradeUpgrade Mellanox InfiniBand switch OSInfiniBand infrastructure maintenance
NVLink Switch Firmware UpgradeUpgrade NVLink switch firmwareNVLink infrastructure maintenance
Device ReprovisioningReprovision network devicesDevice lifecycle management
BMC ProvisioningDeprecated: Not for production useServer infrastructure setup

Data Center Management

WorkflowDescriptionPurpose
Connected Host MetadataCollect and manage host connectivity metadataNetwork topology discovery
VPC CreationCreate virtual private cloudsCloud infrastructure deployment
VPC DeletionDelete virtual private cloudsInfrastructure cleanup

InfiniBand Partition Management

WorkflowDescriptionPurpose
IB PKey CreationAllocate an InfiniBand Partition Key on UFM and record it in NautobotTenant isolation on the IB fabric
IB PKey Member AddAttach device interfaces to an existing PKeyGrow a PKey’s membership
IB PKey Member UpdateReconcile a PKey’s membership to a declarative interface listReplace a PKey’s exact member set
IB PKey Member DeleteDetach device interfaces from an existing PKeyShrink a PKey’s membership

Network Discovery

WorkflowDescriptionPurpose
Port LLDP InfoCollect LLDP information from network portsNetwork topology mapping
Infiniband Get Unhealthy PortsIdentify unhealthy InfiniBand portsNetwork health monitoring

Workflow Architecture

All workflows in Config Manager Temporal follow a consistent architecture:

  • Input validation: Structured input models with type checking
  • Stage-based execution: Multi-stage workflows with approval gates
  • Error handling: Comprehensive error management and retry policies
  • Result archiving: Persistent storage of workflow results
  • Observability: Full execution history and monitoring

Common Input Patterns

Many workflows share common input patterns:

  • Device ID: Identifier for target network device
  • Site Name: Site identifier for multi-device operations
  • User Context: User information for audit trails
  • Configuration Data: Device-specific configuration parameters

Getting Started

  1. Choose the appropriate workflow from the list above
  2. Review the input requirements for your selected workflow
  3. Prepare the input data according to the workflow specification
  4. Execute the workflow using the API or UI
  5. Monitor progress through the workflow stages
  6. Review results and output data

For detailed information about each workflow, click the workflow names in the tables above.