068a476f39
Per-instance ProcessSupervisor: tokio child spawn with proper arg list
(fixes Go's naive space-splitting), graceful SIGTERM with 30s budget
then force kill, monitor task classifying ordered-stop vs crash (exit
code captured), watch-channel state observable everywhere. Instance cmd
channel live on corrosion.{license}.{instance}.cmd (start/stop/restart/
status) with state events pushed on {instance}.status (keep-latest
semantics, documented). Heartbeats now carry live process state +
uptime per instance. Crate restructured lib+bin for integration tests.
Verified: 5 integration tests with real OS processes (lifecycle, crash
exit-code, restart recovery, unmanaged rejection, clean spawn failure)
+ live-NATS contract test (request-reply roundtrips, double-start
rejection, push events, heartbeat state) — all green.
Known limitation (documented): no PID adoption yet — agent restart
orphans a running game process to 'stopped' until panel restart.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2.0 KiB
2.0 KiB
Corrosion Host Agent
Rust rewrite of the Go companion agent (companion-agent/, retained as the
behavior reference until parity). One agent per machine supervises every game
instance on that host — Rust, Conan Exiles, Soulmask, Dune: Awakening.
- Wire protocol: see PROTOCOL.md (v2, instance-scoped subjects)
- Config: see agent.example.toml
Status — Phase 0
- Multi-instance TOML config + env overrides (
CORROSION_LICENSE_ID,CORROSION_NATS_URL,CORROSION_NATS_TOKEN) - NATS connection (infinite reconnect, capped backoff, 30s ping, offline send-buffering,
tls://support) - Host heartbeat with real telemetry (sysinfo: CPU, memory, disks) — no fabricated values
- Connectivity prober (outbound TCP, periodic + on-demand)
- Host command channel (
ping,probe,sysinfo) - Graceful shutdown (cancellation token, going-offline beacon, NATS flush)
- Phase 1a: process supervision — per-instance start/stop/restart/status over
{instance}.cmdrequest-reply, push state events on{instance}.status, crash detection with exit codes, live state in heartbeats (integration-tested with real processes + live-NATS contract test) - Phase 1b: RCON trait (WebRCON rust / TCP conan+soulmask), SteamCMD, jailed file manager
- Phase 2: Dune Docker adapter (compose lifecycle, RabbitMQ bus, Postgres admin)
- Phase 3: signed self-update (enforced ed25519 — release gate), service install, supervisor split
Build
cargo build --release # native
cargo build --release --target x86_64-unknown-linux-gnu # linux deploy target
cargo build --release --target x86_64-pc-windows-msvc # windows (cargo-xwin on non-Windows)
Run
corrosion-host-agent --config ./agent.toml # foreground
corrosion-host-agent --config ./agent.toml check # validate config only
corrosion-host-agent version # semver + git hash + build ts