ananke/internal/service/daemon_additional_test.go

package service

import (
	"context"
	"io"
	"log"
	"os"
	"path/filepath"
	"strings"
	"testing"
	"time"

	"scm.bstein.dev/bstein/ananke/internal/cluster"
	"scm.bstein.dev/bstein/ananke/internal/config"
	"scm.bstein.dev/bstein/ananke/internal/execx"
	"scm.bstein.dev/bstein/ananke/internal/metrics"
	"scm.bstein.dev/bstein/ananke/internal/state"
	"scm.bstein.dev/bstein/ananke/internal/ups"
)

type daemonFakeProvider struct {
	samples []ups.Sample
	errs    []error
	idx     int
}

// Read runs one orchestration or CLI step.
// Signature: (p *daemonFakeProvider) Read(ctx context.Context) (ups.Sample, error).
// Why: daemon tests need deterministic telemetry/error sequencing without real UPS I/O.
func (p *daemonFakeProvider) Read(_ context.Context) (ups.Sample, error) {
	if p.idx < len(p.errs) && p.errs[p.idx] != nil {
		err := p.errs[p.idx]
		p.idx++
		return ups.Sample{}, err
	}
	if p.idx < len(p.samples) {
		s := p.samples[p.idx]
		p.idx++
		return s, nil
	}
	if len(p.samples) > 0 {
		return p.samples[len(p.samples)-1], nil
	}
	return ups.Sample{}, context.DeadlineExceeded
}

// newDaemonTestOrchestrator runs one orchestration or CLI step.
// Signature: newDaemonTestOrchestrator(t *testing.T, stateDir string) *cluster.Orchestrator.
// Why: daemon tests share a minimal dry-run orchestrator fixture to avoid duplication.
func newDaemonTestOrchestrator(t *testing.T, stateDir string) *cluster.Orchestrator {
	t.Helper()
	cfg := config.Config{
		ControlPlanes:   []string{"titan-0a"},
		Workers:         []string{"titan-22"},
		SSHUser:         "atlas",
		SSHPort:         2277,
		SSHManagedNodes: []string{"titan-0a", "titan-22"},
		SSHNodeHosts: map[string]string{
			"titan-0a": "192.168.22.11",
			"titan-22": "192.168.22.22",
		},
		State: config.State{
			Dir:            stateDir,
			ReportsDir:     filepath.Join(stateDir, "reports"),
			RunHistoryPath: filepath.Join(stateDir, "runs.json"),
			LockPath:       filepath.Join(stateDir, "ananke.lock"),
			IntentPath:     filepath.Join(stateDir, "intent.json"),
		},
		Shutdown: config.Shutdown{
			EmergencySkipDrain: true,
			EmergencySkipEtcd:  true,
		},
	}
	return cluster.New(
		cfg,
		&execx.Runner{DryRun: true, Logger: log.New(io.Discard, "", 0)},
		state.New(filepath.Join(stateDir, "runs.json")),
		log.New(io.Discard, "", 0),
	)
}

// TestDaemonRunTriggersShutdownOnLowBattery runs one orchestration or CLI step.
// Signature: TestDaemonRunTriggersShutdownOnLowBattery(t *testing.T).
// Why: covers main daemon loop path that triggers shutdown after debounce threshold.
func TestDaemonRunTriggersShutdownOnLowBattery(t *testing.T) {
	stateDir := t.TempDir()
	orch := newDaemonTestOrchestrator(t, stateDir)
	d := &Daemon{
		cfg: config.Config{
			UPS: config.UPS{
				Enabled:             true,
				PollSeconds:         1,
				DebounceCount:       1,
				RuntimeSafetyFactor: 1.0,
			},
			State: config.State{
				IntentPath: filepath.Join(stateDir, "intent.json"),
			},
			Shutdown: config.Shutdown{
				EmergencySkipDrain: true,
				EmergencySkipEtcd:  true,
			},
		},
		orch: orch,
		targets: []Target{
			{
				Name:   "Pyrphoros",
				Target: "pyrphoros@localhost",
				Provider: &daemonFakeProvider{
					samples: []ups.Sample{{OnBattery: true, LowBattery: true, RuntimeSeconds: 30, RawStatus: "OB LB"}},
				},
			},
		},
		log:      log.New(io.Discard, "", 0),
		exporter: metrics.New(),
	}
	ctx, cancel := context.WithTimeout(context.Background(), 4*time.Second)
	defer cancel()
	if err := d.Run(ctx); err != nil {
		t.Fatalf("expected daemon to trigger and complete shutdown, got %v", err)
	}
}

// TestDaemonRunTriggersShutdownOnTelemetryTimeout runs one orchestration or CLI step.
// Signature: TestDaemonRunTriggersShutdownOnTelemetryTimeout(t *testing.T).
// Why: covers telemetry-timeout trigger path while UPS remains on-battery.
func TestDaemonRunTriggersShutdownOnTelemetryTimeout(t *testing.T) {
	stateDir := t.TempDir()
	orch := newDaemonTestOrchestrator(t, stateDir)
	d := &Daemon{
		cfg: config.Config{
			UPS: config.UPS{
				Enabled:                 true,
				PollSeconds:             1,
				DebounceCount:           3,
				RuntimeSafetyFactor:     1.0,
				TelemetryTimeoutSeconds: 1,
			},
			State: config.State{
				IntentPath: filepath.Join(stateDir, "intent.json"),
			},
			Shutdown: config.Shutdown{
				EmergencySkipDrain: true,
				EmergencySkipEtcd:  true,
			},
		},
		orch: orch,
		targets: []Target{
			{
				Name:   "Statera",
				Target: "statera@localhost",
				Provider: &daemonFakeProvider{
					samples: []ups.Sample{{OnBattery: true, LowBattery: false, RuntimeSeconds: 9999, RawStatus: "OB"}},
					errs:    []error{nil, context.DeadlineExceeded, context.DeadlineExceeded, context.DeadlineExceeded},
				},
			},
		},
		log:      log.New(io.Discard, "", 0),
		exporter: metrics.New(),
	}
	ctx, cancel := context.WithTimeout(context.Background(), 6*time.Second)
	defer cancel()
	if err := d.Run(ctx); err != nil {
		t.Fatalf("expected telemetry-timeout shutdown path to complete, got %v", err)
	}
}

// TestDaemonRunTriggersShutdownAfterOnBatteryGrace runs one orchestration or CLI step.
// Signature: TestDaemonRunTriggersShutdownAfterOnBatteryGrace(t *testing.T).
// Why: covers the sustained-on-battery trigger so short runtime estimates are not
// the only path to a graceful shutdown during abrupt power loss.
func TestDaemonRunTriggersShutdownAfterOnBatteryGrace(t *testing.T) {
	stateDir := t.TempDir()
	orch := newDaemonTestOrchestrator(t, stateDir)
	d := &Daemon{
		cfg: config.Config{
			UPS: config.UPS{
				Enabled:               true,
				PollSeconds:           1,
				DebounceCount:         1,
				RuntimeSafetyFactor:   1.0,
				OnBatteryGraceSeconds: 1,
			},
			State: config.State{
				IntentPath: filepath.Join(stateDir, "intent.json"),
			},
			Shutdown: config.Shutdown{
				EmergencySkipDrain: true,
				EmergencySkipEtcd:  true,
			},
		},
		orch: orch,
		targets: []Target{
			{
				Name:   "Pyrphoros",
				Target: "pyrphoros@localhost",
				Provider: &daemonFakeProvider{
					samples: []ups.Sample{{OnBattery: true, LowBattery: false, RuntimeSeconds: 9999, RawStatus: "OB"}},
				},
			},
		},
		log:      log.New(io.Discard, "", 0),
		exporter: metrics.New(),
	}
	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
	defer cancel()
	if err := d.Run(ctx); err != nil {
		t.Fatalf("expected sustained-on-battery shutdown path to complete, got %v", err)
	}
}

// TestForwardShutdownSucceedsWithSSHShim runs one orchestration or CLI step.
// Signature: TestForwardShutdownSucceedsWithSSHShim(t *testing.T).
// Why: covers forward-shutdown SSH execution path.
func TestForwardShutdownSucceedsWithSSHShim(t *testing.T) {
	tmp := t.TempDir()
	sshPath := filepath.Join(tmp, "ssh")
	script := `#!/usr/bin/env bash
set -euo pipefail
echo forwarded
`
	if err := os.WriteFile(sshPath, []byte(script), 0o755); err != nil {
		t.Fatalf("write fake ssh: %v", err)
	}
	t.Setenv("PATH", tmp+":"+os.Getenv("PATH"))

	d := &Daemon{
		cfg: config.Config{
			SSHUser: "atlas",
			SSHPort: 2277,
			Coordination: config.Coordination{
				ForwardShutdownHost:   "titan-db",
				ForwardShutdownConfig: "/etc/ananke/ananke.yaml",
				CommandTimeoutSeconds: 5,
			},
		},
		log: log.New(io.Discard, "", 0),
	}
	if err := d.forwardShutdown(context.Background(), "test-forward"); err != nil {
		t.Fatalf("forwardShutdown failed: %v", err)
	}
}

// TestForwardShutdownFailsWhenSSHFailsAndNoRecovery runs one orchestration or CLI step.
// Signature: TestForwardShutdownFailsWhenSSHFailsAndNoRecovery(t *testing.T).
// Why: covers forwarded shutdown error propagation branch.
func TestForwardShutdownFailsWhenSSHFailsAndNoRecovery(t *testing.T) {
	tmp := t.TempDir()
	sshPath := filepath.Join(tmp, "ssh")
	script := `#!/usr/bin/env bash
set -euo pipefail
echo "permission denied" >&2
exit 255
`
	if err := os.WriteFile(sshPath, []byte(script), 0o755); err != nil {
		t.Fatalf("write fake ssh: %v", err)
	}
	t.Setenv("PATH", tmp+":"+os.Getenv("PATH"))

	d := &Daemon{
		cfg: config.Config{
			SSHUser: "atlas",
			SSHPort: 2277,
			Coordination: config.Coordination{
				ForwardShutdownHost:   "titan-db",
				ForwardShutdownConfig: "/etc/ananke/ananke.yaml",
				CommandTimeoutSeconds: 5,
			},
		},
		log: log.New(io.Discard, "", 0),
	}
	err := d.forwardShutdown(context.Background(), "test-fail")
	if err == nil {
		t.Fatalf("expected forwardShutdown error")
	}
	if !strings.Contains(strings.ToLower(err.Error()), "forward shutdown via ssh failed") {
		t.Fatalf("unexpected error: %v", err)
	}
}

// TestStartMetricsServerSuccess runs one orchestration or CLI step.
// Signature: TestStartMetricsServerSuccess(t *testing.T).
// Why: covers successful metrics server startup branch.
func TestStartMetricsServerSuccess(t *testing.T) {
	d := &Daemon{
		cfg: config.Config{
			Metrics: config.Metrics{
				Enabled:  true,
				BindAddr: "127.0.0.1:0",
				Path:     "/metrics",
			},
		},
		log:      log.New(io.Discard, "", 0),
		exporter: metrics.New(),
	}
	if err := d.startMetricsServer(); err != nil {
		t.Fatalf("startMetricsServer failed: %v", err)
	}
}