HyperFleet Scripts

CLI utilities for managing HyperFleet clusters, node pools, databases, and Kubernetes resources.

Quick Start

hf.config.sh bootstrap my-env # Interactive setup: context, API, port-forwards, DB
hf.config.sh doctor # Check which scripts are ready to use

Script Groups

Prefix	What it manages
hf.cluster.*	Cluster API objects — create, inspect, patch, watch conditions and statuses
hf.nodepool.*	NodePool API objects — create, inspect, patch, watch conditions and statuses
hf.db.*	Direct database access — queries, status records, cleanup
hf.kube.*	Kubernetes helpers — context, port-forward, in-cluster curl, debug pods
hf.logs.*	Log tailing — pods and adapter-filtered logs
hf.maestro.*	Maestro resource management via maestro-cli
hf.pubsub.*	GCP Pub/Sub — list topics/subscriptions, publish change events
hf.rabbitmq.*	RabbitMQ — publish change events via HTTP management API
hf.repos.*	GitHub repository status overview
hf.config.*	Configuration management — environments, bootstrap, doctor

Config System

All scripts use a file-based configuration system with environment variable overrides:

• Config location: ~/.config/hf/ (one file per setting)
• Precedence: Environment variables > config files > defaults
• Shared state: Settings persist across script invocations
• Property registry: All config keys are declared once in HF_CONFIG_REGISTRY in hf.lib.sh
• Dependency tracking: Each script declares its required config via hf_require_config

Managed via hf.config.sh

hf.config.sh # Show help, environments, and active config
hf.config.sh show # Show current configuration
hf.config.sh show my-env # Show config with environment overrides highlighted
hf.config.sh set api-url <url># Set API URL
hf.config.sh clear token # Clear auth token
hf.config.sh clear all # Reset everything
hf.config.sh doctor # Check per-script readiness
hf.config.sh bootstrap [env] # Interactive environment setup
hf.config.sh env list # List environments
hf.config.sh env activate <name># Activate an environment

Settings are grouped into sections in the registry:

hyperfleet: api-url, api-version, token, context, namespace, gcp-project, cluster-id, cluster-name, nodepool-id

maestro: maestro-consumer, maestro-http-endpoint, maestro-grpc-endpoint, maestro-namespace

portforward: pf-api-port, pf-pg-port, pf-maestro-http-port, pf-maestro-http-remote-port, pf-maestro-grpc-port

database: db-host, db-port, db-name, db-user, db-password

Adding a new config property

1. Add one line to HF_CONFIG_REGISTRY in hf.lib.sh
2. Add a HF_*_FILE variable and _hf_load call in the same file
3. Add a setter function if scripts need to set it programmatically
4. Scripts that need it add the key to their hf_require_config call

No other files need to change -- hf.config.sh, environments, and doctor all derive from the registry.

Scripts Overview

Cluster Management

Script	Description
hf.cluster.create.sh	Create a new cluster via API
hf.cluster.delete.sh	Delete cluster by ID
hf.cluster.get.sh	Get cluster details by ID
hf.cluster.search.sh	Search clusters by name and set as current
hf.cluster.list.sh	List all clusters
hf.cluster.table.sh	Display clusters in table format
hf.cluster.id.sh	Show current cluster ID
hf.cluster.patch.sh	Increment counter field in cluster spec or labels
hf.cluster.conditions.sh	Show cluster conditions (-w for watch)
hf.cluster.conditions.table.sh	Show cluster conditions in table format
hf.cluster.statuses.sh	Show cluster adapter statuses (-w for watch)
hf.cluster.statuses.table.sh	Display cluster adapter statuses in table format (Available and Finalized)
hf.cluster.adapter.post.status.sh	Post adapter status for current cluster

NodePool Management

Script	Description
hf.nodepool.create.sh	Create one or more node pools under the current cluster (name [count] [instance-type])
hf.nodepool.delete.sh	Delete a node pool
hf.nodepool.get.sh	Get node pool details
hf.nodepool.list.sh	List node pools for the current cluster
hf.nodepool.search.sh	Search for nodepool by name and set as current
hf.nodepool.table.sh	Display node pools in table format
hf.nodepool.patch.sh	Increment counter field in nodepool spec or labels
hf.nodepool.conditions.sh	Show node pool conditions (-w for watch)
hf.nodepool.conditions.table.sh	Show node pool conditions in table format
hf.nodepool.statuses.sh	Show node pool adapter statuses (-w for watch)
hf.nodepool.statuses.table.sh	Display node pool adapter statuses in table format (Available and Finalized)
hf.nodepool.adapter.post.status.sh	Post adapter status for current node pool

Database Operations

Script	Description
hf.db.config.sh	Interactive database configuration
hf.db.query.sh	Execute SQL queries or files
hf.db.delete.sh	Delete database records
hf.db.statuses.sh	Query cluster status table
hf.db.delete.all.sh	Delete all records from adapter_statuses, node_pools, and clusters tables
hf.db.statuses.delete.sh	Delete status records

Kubernetes Helpers

Script	Description
hf.kube.context.sh	Select and save kubectl context/namespace
hf.kube.curl.sh	Run curl from inside a cluster pod (see below)
hf.kube.debug.pod.sh	Create debug pod cloned from a deployment (see below)
hf.kube.port.forward.sh	Port forward to services/pods

hf.kube.curl.sh

Runs curl from inside a pod in the current Kubernetes cluster, useful for reaching cluster-internal services. The pod (hf-kcurl) is reused across invocations and auto-terminates after 5 minutes via sleep 300, so the first call pays the pod startup cost but subsequent calls are fast. Only curl output is written to stdout, making the script safe to use in pipelines and other scripts.

# Simple GET
hf.kube.curl.sh http://my-service.ns.svc:8080/health
# POST with inline data
hf.kube.curl.sh -X POST -H "Content-Type: application/json" -d '{"key":"val"}' http://my-service.ns.svc:8080/api
# POST with file body
hf.kube.curl.sh -X POST -H "Content-Type: application/json" -f payload.json http://my-service.ns.svc:8080/api
# Save response to file
hf.kube.curl.sh -o response.json http://my-service.ns.svc:8080/api
# Use in a pipeline
hf.kube.curl.sh http://my-service.ns.svc:8080/api | jq '.items[]'

hf.kube.debug.pod.sh

Creates a debug pod by cloning the pod template from an existing deployment. The debug pod runs with the same image, environment variables, volumes, and service account as the original deployment, but replaces the entrypoint with sleep infinity and removes all health probes. This lets you exec into a shell with the exact same runtime context as the target workload — useful for debugging configuration, network connectivity, or permissions issues.

The deployment is matched by partial name, so you don't need to type the full name.

# Clone a deployment's pod template into a debug pod and exec into it
hf.kube.debug.pod.sh my-app
# Specify a namespace
hf.kube.debug.pod.sh my-app staging
# Clean up when done
kubectl delete pod my-app-debug-<timestamp> -n default

Logs

Script	Description
hf.logs.sh	Tail pod logs with context
hf.logs.adapter.sh	Tail adapter pod logs filtered by cluster ID

Maestro

Maestro scripts require maestro-cli to be compiled and available on your PATH from openshift-hyperfleet/maestro-cli.

Script	Description
hf.maestro.list.sh	List maestro resources
hf.maestro.get.sh	Get a maestro resource by name (auto-selects if one match; interactive menu if multiple)
hf.maestro.delete.sh	Delete a maestro resource by name (interactive selection if no name given)
hf.maestro.bundles.sh	List maestro resource bundles
hf.maestro.consumers.sh	List maestro consumers
hf.maestro.tui.sh	Launch maestro-cli TUI

Pub/Sub and Messaging

Script	Description
hf.pubsub.list.sh	List Pub/Sub topics and subscriptions, with optional filter
hf.pubsub.publish.cluster.change.sh	Publish cluster change event to a Pub/Sub topic
hf.pubsub.publish.nodepool.change.sh	Publish nodepool change event to a Pub/Sub topic
hf.rabbitmq.publish.cluster.change.sh	Publish cluster change event to a RabbitMQ exchange via HTTP management API

Other Utilities

Script	Description
hf.table.sh	Combined table of all clusters with their nodepools, showing ready status and adapter counts
hf.workflow.sh	End-to-end test workflow: create cluster+nodepool, cycle adapter statuses and patches through multiple generations, then delete
hf.workflow.api-only.sh	API-only workflow: create cluster+nodepool, post adapter statuses, and clean up without real adapters
hf.repos.sh	Show status table for all HyperFleet GitHub repositories
hf.lib.sh	Shared library (config registry, API helpers, logging, Kubernetes wrappers)

Common Patterns

Bootstrap a new environment:

hf.config.sh bootstrap dev

Check what's ready:

hf.config.sh doctor

Search and set current cluster:

hf.cluster.search.sh my-cluster-name

Query cluster details (uses saved cluster-id):

hf.cluster.get.sh

Create and manage node pools:

hf.nodepool.create.sh my-pool 3 m5.2xlarge
hf.nodepool.list.sh
hf.nodepool.table.sh
hf.nodepool.conditions.sh -w

Post adapter status manually:

hf.cluster.adapter.post.status.sh validator True 1
hf.nodepool.adapter.post.status.sh provisioner True 1

Switch environments:

hf.config.sh env list
hf.config.sh env activate staging

Configure kubectl context:

hf.kube.context.sh select

Run database query:

hf.db.query.sh "SELECT * FROM clusters LIMIT 10"
hf.db.query.sh -f schema.sql

Architecture

• hf.lib.sh: Core library providing config registry, config loading, API helpers, logging, and Kubernetes wrappers
• Config registry: HF_CONFIG_REGISTRY in hf.lib.sh is the single source of truth for all config properties (section, key, default, sensitivity)
• Dependency declarations: Each script declares hf_require_config <keys> -- used at runtime for validation and by doctor for readiness scanning
• File-based state: Each config value stored separately in ~/.config/hf/ for easy inspection and editing
• Environment profiles: Named environments stored as <env>.<property> files in the config directory
• Environment override: Set HF_API_URL, HF_TOKEN, etc. to override file config
• Consistent interface: All scripts source hf.lib.sh for unified behavior and error handling