🌱 Before using the ssh Rescue Client, ensure the rescue system is active. Log reboot events

pkg/services/baremetal/host/host.go

@@ -592,46 +592,18 @@ func (s *Service) actionRegistering(_ context.Context) actionResult {
 	}
 	sshClient := s.scope.SSHClientFactory.NewClient(in)
 
-	// Check hostname with sshClient
-	out := sshClient.GetHostName()
-	hostName := trimLineBreak(out.StdOut)
-	if hostName != rescue {
-		// give the reboot some time until it takes effect
-		if s.hasJustRebooted() {
-			return actionContinue{delay: 2 * time.Second}
-		}
-
-		isSSHTimeoutError, isSSHConnectionRefusedError, err := s.analyzeSSHOutputRegistering(out)
-		if err != nil {
-			// This can happen if the bare-metal server was taken by another mgt-cluster.
-			// Check in https://robot.hetzner.com/server for the "History" of the server.
-			return actionError{err: fmt.Errorf("failed to handle incomplete boot - registering: %w", err)}
-		}
-
-		failed, err := s.handleIncompleteBoot(true, isSSHTimeoutError, isSSHConnectionRefusedError)
-		if failed {
-			return s.recordActionFailure(infrav1.PermanentError, err.Error())
-		}
-		if err != nil {
-			return actionError{err: fmt.Errorf(errMsgFailedHandlingIncompleteBoot, err)}
-		}
-		timeSinceReboot := "unknown"
-		if s.scope.HetznerBareMetalHost.Spec.Status.LastUpdated != nil {
-			timeSinceReboot = time.Since(s.scope.HetznerBareMetalHost.Spec.Status.LastUpdated.Time).String()
-		}
-
-		s.scope.Logger.Info("Could not reach rescue system. Will retry some seconds later.", "out", out.String(), "hostName", hostName,
-			"isSSHTimeoutError", isSSHTimeoutError, "isSSHConnectionRefusedError", isSSHConnectionRefusedError, "timeSinceReboot", timeSinceReboot)
-		return actionContinue{delay: 10 * time.Second}
+	ok, res := s.validateRescueSystemIsActive(sshClient)
+	if !ok {
+		return res
 	}
-
 	output := sshClient.GetHardwareDetailsDebug()
 	if output.Err != nil {
 		return actionError{err: fmt.Errorf("failed to obtain hardware for debugging: %w", output.Err)}
 	}
+
 	msg := fmt.Sprintf("%s\n\n", output.StdOut)
-	if out.StdErr != "" {
-		msg += fmt.Sprintf("stderr:\n%s\n\n", out.StdErr)
+	if output.StdErr != "" {
+		msg += fmt.Sprintf("stderr:\n%s\n\n", output.StdErr)
 	}
 	record.Eventf(s.scope.HetznerBareMetalHost, "GetHardwareDetails", msg)
 
@@ -711,6 +683,42 @@ func (s *Service) actionRegistering(_ context.Context) actionResult {
 	return actionComplete{}
 }
 
+func (s *Service) validateRescueSystemIsActive(sshClient sshclient.Client) (ok bool, ar actionResult) {
+	// Check hostname with sshClient
+	out := sshClient.GetHostName()
+	hostName := trimLineBreak(out.StdOut)
+	if hostName == rescue {
+		return true, actionContinue{}
+	}
+	// give the reboot some time until it takes effect
+	if s.hasJustRebooted() {
+		return false, actionContinue{delay: 2 * time.Second}
+	}
+
+	isSSHTimeoutError, isSSHConnectionRefusedError, err := s.analyzeSSHOutputRegistering(out)
+	if err != nil {
+		// This can happen if the bare-metal server was taken by another mgt-cluster.
+		// Check in https://robot.hetzner.com/server for the "History" of the server.
+		return false, actionError{err: fmt.Errorf("failed to handle incomplete boot - registering: %w", err)}
+	}
+
+	failed, err := s.handleIncompleteBoot(true, isSSHTimeoutError, isSSHConnectionRefusedError)
+	if failed {
+		return false, s.recordActionFailure(infrav1.PermanentError, err.Error())
+	}
+	if err != nil {
+		return false, actionError{err: fmt.Errorf(errMsgFailedHandlingIncompleteBoot, err)}
+	}
+	timeSinceReboot := "unknown"
+	if s.scope.HetznerBareMetalHost.Spec.Status.LastUpdated != nil {
+		timeSinceReboot = time.Since(s.scope.HetznerBareMetalHost.Spec.Status.LastUpdated.Time).String()
+	}
+
+	s.scope.Logger.Info("Could not reach rescue system. Will retry some seconds later.", "out", out.String(), "hostName", hostName,
+		"isSSHTimeoutError", isSSHTimeoutError, "isSSHConnectionRefusedError", isSSHConnectionRefusedError, "timeSinceReboot", timeSinceReboot)
+	return false, actionContinue{delay: 10 * time.Second}
+}
+
 func validateRootDeviceWwnsAreSubsetOfExistingWwns(rootDeviceHints *infrav1.RootDeviceHints, storageDevices []infrav1.Storage) error {
 	knownWWNs := make([]string, 0, len(storageDevices))
 	for _, sd := range storageDevices {
@@ -1097,6 +1105,11 @@ func (s *Service) actionPreProvisioning(ctx context.Context) actionResult {
 	}
 	sshClient := s.scope.SSHClientFactory.NewClient(in)
 
+	ok, res := s.validateRescueSystemIsActive(sshClient)
+	if !ok {
+		return res
+	}
+
 	exitStatus, output, err := sshClient.ExecutePreProvisionCommand(ctx, s.scope.PreProvisionCommand)
 	if err != nil {
 		return actionError{err: fmt.Errorf("failed to execute pre-provision command: %w", err)}
@@ -1128,6 +1141,11 @@ func (s *Service) actionImageInstalling(ctx context.Context) actionResult {
 	}
 	sshClient := s.scope.SSHClientFactory.NewClient(in)
 
+	ok, res := s.validateRescueSystemIsActive(sshClient)
+	if !ok {
+		return res
+	}
+
 	state, err := sshClient.GetInstallImageState()
 	if err != nil {
 		return actionError{err: fmt.Errorf("failed to get state of installimage processes: %w", err)}