agent: command.update handler + updater package (Linux + Windows)

cmd/agent/main.go

@@ -148,6 +148,7 @@ func run() error {
 		resticBin:                 resticBin,
 		resticVer:                 snap.ResticVersion,
 		resticSupportsNoOwnership: resticSupportsNoOwnership,
+		serverURL:                 cfg.ServerURL,
 		secrets:                   sec,
 		scheduler:                 scheduler.New(),
 	}
@@ -214,6 +215,7 @@ type dispatcher struct {
 	resticBin                 string
 	resticVer                 string // e.g. "0.17.1"; empty if restic isn't installed yet
 	resticSupportsNoOwnership bool   // captured at startup from `restic restore --help`
+	serverURL                 string // base URL of the server (used by the self-update fetch)
 	secrets                   *secrets.Store
 	scheduler                 *scheduler.Scheduler
 
@@ -395,10 +397,12 @@ func (d *dispatcher) handle(ctx context.Context, env api.Envelope, tx wsclient.S
 				"up_kbps", up, "down_kbps", down)
 		}
 
-	case api.MsgAgentUpdateAvail:
-		var p api.AgentUpdateAvailablePayload
-		_ = env.UnmarshalPayload(&p)
-		slog.Info("ws agent: update available", "version", p.LatestVersion, "url", p.PackageURL)
+	case api.MsgCommandUpdate:
+		var p api.CommandUpdatePayload
+		if err := env.UnmarshalPayload(&p); err != nil {
+			return fmt.Errorf("command.update: %w", err)
+		}
+		go d.runUpdate(ctx, p, tx)
 
 	default:
 		slog.Debug("ws agent: ignored message", "type", env.Type)

cmd/agent/update_dispatch.go

@@ -0,0 +1,65 @@
+package main
+
+import (
+	"context"
+	"fmt"
+	"log/slog"
+	"time"
+
+	"gitea.dcglab.co.uk/steve/restic-manager/internal/agent/updater"
+	"gitea.dcglab.co.uk/steve/restic-manager/internal/agent/wsclient"
+	"gitea.dcglab.co.uk/steve/restic-manager/internal/api"
+)
+
+// runUpdate handles a server-dispatched command.update. It logs progress
+// via log.stream so the live job page captures pre-restart state, then
+// calls the platform updater. On Linux the updater calls os.Exit; on
+// Windows it spawns a detached helper and returns, with the agent then
+// exiting.
+//
+// The terminal job state is set by the server, not the agent: success
+// is "agent re-hellos with matching version" rather than anything the
+// agent itself can assert. The only `job.finished` we send from here is
+// on the failure path, before any restart attempt.
+func (d *dispatcher) runUpdate(ctx context.Context, p api.CommandUpdatePayload, tx wsclient.Sender) {
+	logf := func(format string, args ...any) {
+		line := fmt.Sprintf(format, args...)
+		slog.Info("ws agent: update: " + line)
+		env, err := api.Marshal(api.MsgLogStream, "", api.LogStreamLine{
+			JobID:   p.JobID,
+			TS:      time.Now().UTC(),
+			Stream:  api.LogStdout,
+			Payload: line,
+		})
+		if err == nil {
+			_ = tx.Send(env)
+		}
+	}
+
+	startedEnv, err := api.Marshal(api.MsgJobStarted, "", api.JobStartedPayload{
+		JobID:     p.JobID,
+		Kind:      api.JobUpdate,
+		StartedAt: time.Now().UTC(),
+	})
+	if err == nil {
+		_ = tx.Send(startedEnv)
+	}
+
+	logf("fetching new binary from %s", d.serverURL)
+	if err := updater.Update(ctx, d.serverURL); err != nil {
+		logf("update failed: %v", err)
+		finishedEnv, mErr := api.Marshal(api.MsgJobFinished, "", api.JobFinishedPayload{
+			JobID:      p.JobID,
+			Status:     api.JobFailed,
+			FinishedAt: time.Now().UTC(),
+			Error:      err.Error(),
+		})
+		if mErr == nil {
+			_ = tx.Send(finishedEnv)
+		}
+		return
+	}
+	// Unreachable on Linux (Update calls os.Exit). On Windows control
+	// returns here while the detached helper does the swap-and-restart;
+	// the agent then exits cleanly so SCM hands off.
+}