k3s-io/k3s · error

failed to unmarshal tailscale output: %v

Error message

failed to unmarshal tailscale output: %v

What it means

The stdout of `tailscale status --json` is unmarshalled into TailscaleOutput{TailscaleIPs, BackendState}. If that output is not valid JSON for the shape - the CLI printed an error or warning instead, the output was empty, or a version changed the payload - json.Unmarshal fails and the error is wrapped here.

Source

Thrown at pkg/vpn/vpn.go:171

		return nil
	}

	return errors.New("Requested VPN: " + authInfo.Name + " is not supported. We currently only support tailscale")
}

// getTailscaleInfo returns the IPs of the interface
func getTailscaleInfo() (*Info, error) {
	output, err := util.ExecCommand("tailscale", []string{"status", "--json"})
	if err != nil {
		return nil, fmt.Errorf("failed to run tailscale status --json: %v", err)
	}

	logrus.Debugf("Output from tailscale status --json: %v", output)

	var tailscaleOutput TailscaleOutput
	err = json.Unmarshal([]byte(output), &tailscaleOutput)
	if err != nil {
		return nil, fmt.Errorf("failed to unmarshal tailscale output: %v", err)
	}

	// Errors are ignored because the interface might not have ipv4 or ipv6 addresses (that's the only possible error)
	ipv4Address, _ := util.GetFirst4String(tailscaleOutput.TailscaleIPs)
	ipv6Address, _ := util.GetFirst6String(tailscaleOutput.TailscaleIPs)

	return &Info{BackendState: tailscaleOutput.BackendState, IPv4Address: net.ParseIP(ipv4Address), IPv6Address: net.ParseIP(ipv6Address), NodeID: "", ProviderName: "tailscale", Interface: tailscaleIf}, nil
}

// get Tailscale advertised route list
func GetAdvertisedRoutes() ([]netip.Prefix, error) {
	output, err := util.ExecCommand("tailscale", []string{"debug", "prefs"})
	if err != nil {
		return nil, fmt.Errorf("failed to run tailscale debug prefs: %v", err)
	}

	logrus.Debugf("Output from tailscale debug prefs: %v", output)

View on GitHub (pinned to 6ba341e396)

Solutions

  1. Run tailscale status --json by hand and pipe through jq . to see what is actually returned
  2. Align the CLI to a known-good version and restart tailscaled after the change
  3. If leading warnings pollute the output, fix the underlying warning (often permissions or stale state) first
Defensive patterns

Strategy: try-catch

Validate before calling

// pre-flight: confirm the CLI emits the JSON shape you expect
out, err := exec.Command("tailscale", "status", "--json").Output()
if err != nil {
	return fmt.Errorf("tailscale status failed: %w", err)
}
var probe struct{ BackendState string `json:"BackendState"` }
if err := json.Unmarshal(out, &probe); err != nil {
	return fmt.Errorf("unexpected tailscale output %q: %w", out, err)
}

Try / catch

info, err := vpn.GetInfo(vpnAuth)
if err != nil {
	if strings.Contains(err.Error(), "failed to unmarshal tailscale output") {
		// run `tailscale status --json | jq .` by hand; usually version drift or
		// non-JSON warnings on stdout - align the CLI version and restart tailscaled
		log.Printf("raw tailscale output needs inspection")
	}
	return nil, err
}

Prevention

When it happens

Trigger: tailscale exits 0 but writes non-JSON to stdout (warnings, banners, error text); an empty output; a CLI version whose `status --json` payload differs from the expected fields.

Common situations: Upgrading or downgrading the tailscale CLI without restarting tailscaled; distributions that patch tailscale; environments where the CLI wraps or pollutes stdout.

Related errors


AI-assisted analysis of k3s-io/k3s@6ba341e396 (2026-08-15). Data as JSON: /api/errors/cf38daaa6e0b8a34. Report an issue: GitHub.