restic-manager/cmd/server/main.go

package main

import (
	"context"
	"errors"
	"flag"
	"fmt"
	"log/slog"
	"os"
	"os/signal"
	"path/filepath"
	"syscall"
	"time"

	"gitea.dcglab.co.uk/steve/restic-manager/internal/auth"
	"gitea.dcglab.co.uk/steve/restic-manager/internal/crypto"
	"gitea.dcglab.co.uk/steve/restic-manager/internal/server/config"
	rmhttp "gitea.dcglab.co.uk/steve/restic-manager/internal/server/http"
	"gitea.dcglab.co.uk/steve/restic-manager/internal/server/maintenance"
	"gitea.dcglab.co.uk/steve/restic-manager/internal/server/ui"
	"gitea.dcglab.co.uk/steve/restic-manager/internal/server/ws"
	"gitea.dcglab.co.uk/steve/restic-manager/internal/store"
)

var version = "dev"

func main() {
	if err := run(); err != nil {
		slog.Error("server fatal", "err", err)
		os.Exit(1)
	}
}

func run() error {
	configPath := flag.String("config", "", "path to YAML config (optional; env vars win regardless)")
	showVersion := flag.Bool("version", false, "print version and exit")
	flag.Parse()

	if *showVersion {
		fmt.Println("restic-manager-server", version)
		return nil
	}

	logger := slog.New(slog.NewJSONHandler(os.Stdout, &slog.HandlerOptions{Level: slog.LevelInfo}))
	slog.SetDefault(logger)

	cfg, err := config.Load(*configPath)
	if err != nil {
		return fmt.Errorf("config: %w", err)
	}
	slog.Info("config resolved", "listen", cfg.Listen, "data_dir", cfg.DataDir,
		"cookie_secure", cfg.CookieSecure, "trusted_proxies", cfg.TrustedProxies)

	if err := os.MkdirAll(cfg.DataDir, 0o700); err != nil {
		return fmt.Errorf("ensure data dir: %w", err)
	}

	// Mint or load the encryption key.
	if _, err := os.Stat(cfg.SecretKeyFile); errors.Is(err, os.ErrNotExist) {
		slog.Warn("no secret key found; generating a new one — back this up before continuing",
			"path", cfg.SecretKeyFile)
		if err := crypto.GenerateKeyFile(cfg.SecretKeyFile); err != nil {
			return fmt.Errorf("generate secret key: %w", err)
		}
	}
	keyBytes, err := crypto.LoadKeyFromFile(cfg.SecretKeyFile)
	if err != nil {
		return fmt.Errorf("load secret key: %w", err)
	}
	aead, err := crypto.NewAEAD(keyBytes)
	if err != nil {
		return fmt.Errorf("init AEAD: %w", err)
	}

	dbPath := filepath.Join(cfg.DataDir, "restic-manager.db")
	st, err := store.Open(context.Background(), dbPath)
	if err != nil {
		return fmt.Errorf("open store: %w", err)
	}
	defer func() { _ = st.Close() }()

	hub := ws.NewHub()
	jobHub := ws.NewJobHub()

	renderer, err := ui.New()
	if err != nil {
		return fmt.Errorf("ui: %w", err)
	}

	deps := rmhttp.Deps{
		Cfg:     cfg,
		Store:   st,
		AEAD:    aead,
		Hub:     hub,
		JobHub:  jobHub,
		UI:      renderer,
		Version: version,
	}

	// First-run bootstrap: if the users table is empty, mint a one-time
	// token and print it. /api/bootstrap accepts it to create the first
	// admin user, then becomes a no-op.
	n, err := st.CountUsers(context.Background())
	if err != nil {
		return fmt.Errorf("count users: %w", err)
	}
	if n == 0 {
		token, err := auth.NewToken()
		if err != nil {
			return fmt.Errorf("mint bootstrap token: %w", err)
		}
		deps.BootstrapToken = token
		// Stable, easy-to-grep marker so an operator finds this in
		// scrolling logs without spelunking. Token is shown in plain
		// text exactly once; we hash it into BootstrapToken on the
		// server-side handler.
		fmt.Fprintln(os.Stderr, "================================================================")
		fmt.Fprintln(os.Stderr, "  FIRST RUN — bootstrap token (use within 1 hour, then it's gone):")
		fmt.Fprintln(os.Stderr, "    "+token)
		fmt.Fprintln(os.Stderr, "  POST it to /api/bootstrap with {token, username, password}.")
		fmt.Fprintln(os.Stderr, "================================================================")
	}

	srv := rmhttp.New(deps)

	ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
	defer stop()

	errCh := make(chan error, 1)
	go func() {
		slog.Info("server listening", "addr", cfg.Listen, "version", version)
		errCh <- srv.Start()
	}()

	// Background sweepers:
	//   - sessions/tokens purge: 15 min
	//   - host offline-after-90s mark: every 30s (matches heartbeat
	//     cadence — agent sends every 30s, P1-12)
	purgeTick := time.NewTicker(15 * time.Minute)
	defer purgeTick.Stop()
	offlineTick := time.NewTicker(30 * time.Second)
	defer offlineTick.Stop()
	// Maintenance ticker: drives forget/prune/check on the cadences
	// operators set per-host. Independent of the agent's local cron
	// (which only handles backup schedules). 60s cadence — the cron
	// expressions are minute-grained, so anything finer is wasted
	// work.
	maintenanceTick := time.NewTicker(60 * time.Second)
	defer maintenanceTick.Stop()
	mt := maintenance.New(st)
	go func() {
		for {
			select {
			case <-ctx.Done():
				return
			case <-purgeTick.C:
				if n, err := st.PurgeExpiredSessions(ctx); err == nil && n > 0 {
					slog.Info("purged expired sessions", "n", n)
				}
				if n, err := st.PurgeExpiredEnrollmentTokens(ctx); err == nil && n > 0 {
					slog.Info("purged expired enrollment tokens", "n", n)
				}
			case <-offlineTick.C:
				cutoff := time.Now().Add(-90 * time.Second)
				if n, err := st.MarkHostsOfflineStale(ctx, cutoff); err == nil && n > 0 {
					slog.Info("marked hosts offline (stale heartbeat)", "n", n)
				}
			case <-maintenanceTick.C:
				decisions, err := mt.Decide(ctx, time.Now().UTC())
				if err != nil {
					slog.Warn("maintenance ticker: decide", "err", err)
					continue
				}
				if len(decisions) > 0 {
					slog.Info("maintenance ticker: dispatching", "n", len(decisions))
					srv.DispatchMaintenance(ctx, decisions)
				}
			}
		}
	}()

	select {
	case err := <-errCh:
		if err != nil {
			return fmt.Errorf("listen: %w", err)
		}
	case <-ctx.Done():
		slog.Info("shutting down")
	}

	shutCtx, cancel := context.WithTimeout(context.Background(), 15*time.Second)
	defer cancel()
	if err := srv.Shutdown(shutCtx); err != nil {
		return fmt.Errorf("shutdown: %w", err)
	}
	return nil
}