juicedata/juicefs · error

error GET request: %v

Error message

error GET request: %v

What it means

getRequest issues the request with http.DefaultClient.Do; any transport-level failure — connection refused, DNS failure, timeout (context deadline exceeded), TLS error — is wrapped as 'error GET request: %v'. The request was created successfully but never completed successfully at the network layer.

Source

Thrown at cmd/debug.go:231

		}
	}

	if listenPort == -1 {
		return 0, fmt.Errorf("no valid pprof port found")
	}
	return listenPort, nil
}

func getRequest(url string, timeout time.Duration) ([]byte, error) {
	ctx, cancel := context.WithTimeout(context.Background(), timeout)
	defer cancel()
	req, err := http.NewRequestWithContext(ctx, "GET", url, nil)
	if err != nil {
		return nil, fmt.Errorf("error creating GET request: %v", err)
	}
	resp, err := http.DefaultClient.Do(req)
	if err != nil {
		return nil, fmt.Errorf("error GET request: %v", err)
	}
	if resp.StatusCode != 200 {
		return nil, fmt.Errorf("error GET request, status code %d", resp.StatusCode)
	}

	defer func(body io.ReadCloser) {
		if err := body.Close(); err != nil {
			logger.Errorf("error closing body: %v", err)
		}
	}(resp.Body)
	body, err := io.ReadAll(resp.Body)
	if err != nil {
		return nil, fmt.Errorf("error reading response: %v", err)
	}

	return body, nil
}

View on GitHub (pinned to c9a67b23e8)

Solutions

1. Verify the endpoint manually: `curl -m 3 http://localhost:<port>/debug/pprof/cmdline?debug=1` and compare the failure.
2. If the error is `context deadline exceeded`, increase the timeout argument passed to getRequest / reqAndSaveMetric.
3. If connection refused, confirm the process is still alive and listening on that port (`ss -ltnp | grep <pid>`), then rerun `juicefs debug`.
4. Check proxy environment variables (HTTP_PROXY/HTTPS_PROXY) which http.DefaultClient honors and which can break localhost requests.
5. Test metric URLs (prometheus/grafana endpoints) reachability and DNS from the same host.

Example fix

// before
resp, err := getRequest("http://localhost:6060/debug/pprof/heap?debug=1", 3*time.Second)
// after
for i := 0; i < 3; i++ {
    resp, err = getRequest(url, 3*time.Second)
    if err == nil { break }
    time.Sleep(500 * time.Millisecond)
}

conn, err := net.DialTimeout("tcp", fmt.Sprintf("localhost:%d", port), 2*time.Second)
if err != nil {
    return fmt.Errorf("pprof port %d not reachable: %v", port, err)
}
conn.Close()

var body []byte
var err error
for i := 0; i < 3; i++ {
    body, err = getRequest(url, 5*time.Second)
    if err == nil { break }
    var ne net.Error
    if errors.As(err, &ne) && ne.Timeout() {
        time.Sleep(time.Second) // retry only transient timeouts
        continue
    }
    break // connection refused etc. — don't retry blindly
}

When it happens

Trigger: Called from checkPort (probing localhost:<port>/debug/pprof/cmdline with a 3s timeout) and reqAndSaveMetric (metric/profile URLs) when: the pprof listener is not actually up on that port, the connection is refused/times out, or the metric endpoint host is unreachable.

Common situations: A stale lsof-derived port belongs to a socket that just closed; pprof is listening on IPv6 only and localhost resolves oddly; a 3s timeout is too short on a heavily loaded client (context deadline exceeded); remote metric hosts behind firewalls/VPN drops.

Understand the failure class

Background: 'Something went wrong' / 'Request failed (500)' / 'HTTP error! status: 404' — what failed HTTP requests actually mean and how to find the real cause — this error's family across 28 libraries.

Related errors

• error creating GET request: %v
• error GET request, status code %d
• get range [%v-%v): %s
• bad response status %s
• got %s

AI-assisted analysis of juicedata/juicefs@c9a67b23e8 (2026-09-06). Data as JSON: /api/errors/0664e7e013064b6b. Report an issue: GitHub.