Polaripol CLI quick reference
pol help overview (from the command table)
pol <module> help per-module help
# Security substrate (everything self-generating; skip-if-exists)
pol security setup [dev|prod] [--env-only|--certs-only|--skip-subs]
pol security node-setup [staging|prod] PRF standalone env + .generated
pol security status which credential files exist / legacy
pol security cleanup remove all generated certs+creds
# Compose orchestration (the existing compose family) [alias: c]
# Env tiers per project: rf-node has dev|test|staging|prod(+stateless);
# suite root has dev|staging|prod (test exists only at rf-node level).
pol compose suite|node <action> [--env E] full roles
pol compose engines|dask up|down|ps|logs independent service deploys
pol compose twin <args> via twin-polari-build.sh
pol compose node up backend any single service by name
pol suite … / pol node … shortcuts for the two main roles
pol suite urls staging nip.io URLs
# Swarm orchestration (the isle-mesh STAND-IN) — LIVE
pol swarm init|status|join-token cluster management
pol swarm render|deploy|rm|ps|services <role> stacks (roles: engines|suite|node)
# engines proven E2E; suite/node conflict with their compose twins (refuse)
# v1 inlines generated env values via compose-config; secrets = refinement
# Generated nginx proxies (replaces the sed .template path)
pol proxy render|check|promote|status check = nginx -t in a container
# rf prod render needs POLARI_PROD_DOMAIN exported
# ssh deploys to configured nodes (pol-build/manifests/nodes.yml)
pol deploy nodes|preflight <node>
pol deploy run <node> --role engines|remote-worker|node [--dry-run]
# push branches first — targets pull the PUBLIC github repos
# Isle-mesh mode (future) — refuses; swarm stands in
pol isle
# Service accountability [alias: reg, services]
pol registry list|show <kind>|interconnects|check
# Topology as core-instance data (top-1..8) [alias: top, topo]
pol topology status|graph|validate|diff read views (drift + findings)
pol topology pull|push [file] rows <-> topologies/*.topology.yml
pol topology report [--node <n>] observe what actually runs
pol topology render|apply [--plan] rows -> manifests -> running (parity-gated)
pol topology deploy <package> [--plan] portable-package flow (push+render+apply)
pol topology assign <module> <instance> move a module (rows only)
pol allocate <module|instance> <instance|machine> targeted deploy (swarm placement)
pol swarm join <node> drive a nodes.yml machine into the swarm
# Effective configuration (read-only; nested per-service) [alias: cfg]
pol config show|knobs|env|generated
pol config service <kind> [show|files|knobs|connects]
# Database backend per PRF instance
pol db show|options
pol db use combo|sqlite --role twin primary switching = honest refusal
# PRF feature modules [alias: mod]
pol modules list|deps|selftest <module> enable/disable -> pol topology assign
# Certificates per env tier [alias: certs, ca]
pol cert setup|issue|verify|walkthrough|renew
pol cert prod self-signed | pol cert prod letsencrypt [--dry-run]
pol cert auto-renew install|status|remove (open-source cron + certbot)
# Build pipeline (jinja-script) [alias: b]
pol build render [--topology single|swarm] swarm = bld-5, refuses today
pol build parity semantic compose-config diff
pol build detect [dir] list jinja-annotated files
pol build clean rm rendered jinja-build/
# CA toolkit [alias: certs, ca]
pol cert setup|issue|renew|verify|walkthroughKnobs honored everywhere: LOCAL_IP, POLARI_SUITE_ROOT, and the credential knobs (POLARI_KC_ADMIN_PASS, POLARI_MYSQL_ROOT_PASS, POLARI_KC_DB_PASS, POLARI_PSC_DB_PASS, POLARI_MINIO_ROOT_USER/_PASS, POLARI_MARIADB_ROOT_PASS, POLARI_OBJECTS_DB_PASS, POLARI_KEYDB_PASS).
Gotchas:
- Standalone node and combined suite share container names — run one.
- Combined staging first
up: prf-backend healthcheck can flap on a cold seed and block pol-proxy — re-runpol suite uponce it's healthy.