flipped-aurora/gin-vue-admin · error

向客户端写入流式响应失败: %w

Error message

向客户端写入流式响应失败: %w

What it means

proxyLLMStream copies the upstream stream to the client in 32KB chunks: it reads from res.Body and writes to c.Writer via c.Writer.Write. If any client-facing write fails, it returns fmt.Errorf("向客户端写入流式响应失败: %w", writeErr) — 'failed to write streaming response to client' — wrapping the underlying write error. This usually means the client connection is gone.

Source

Thrown at server/api/v1/system/sys_auto_code.go:201

	copyLLMStreamHeaders(c.Writer.Header(), res.Header)
	if c.Writer.Header().Get("Content-Type") == "" {
		c.Writer.Header().Set("Content-Type", "text/event-stream; charset=utf-8")
	}
	if c.Writer.Header().Get("Cache-Control") == "" {
		c.Writer.Header().Set("Cache-Control", "no-cache")
	}
	c.Writer.Header().Set("Connection", "keep-alive")
	c.Writer.Header().Set("X-Accel-Buffering", "no")
	c.Status(res.StatusCode)
	flusher.Flush()

	buf := make([]byte, 32*1024)
	for {
		n, readErr := res.Body.Read(buf)
		if n > 0 {
			if _, writeErr := c.Writer.Write(buf[:n]); writeErr != nil {
				return fmt.Errorf("向客户端写入流式响应失败: %w", writeErr)
			}
			flusher.Flush()
		}
		if readErr != nil {
			if errors.Is(readErr, io.EOF) {
				return nil
			}
			return fmt.Errorf("读取上游流式响应失败: %w", readErr)
		}
	}
}

func copyLLMStreamHeaders(dst, src http.Header) {
	for _, key := range []string{
		"Content-Type",
		"Cache-Control",
		"Content-Encoding",
		"Content-Language",

View on GitHub (pinned to 3136500ef3)

Solutions

1. Check the wrapped %w error: context canceled / broken pipe means the client left — handle as a normal cancellation, not a server bug.
2. Raise proxy timeouts (nginx proxy_read_timeout, LB idle timeout) to exceed the expected LLM generation duration.
3. Ensure the frontend keeps the request alive and does not abort early; surface cancel handling in the fetch layer.
4. In the handler, detect ctx.Err() (c.Request.Context().Err()) after write failure and return a sentinel so the server logs it as cancellation instead of an error.

Example fix

// before
if _, writeErr := c.Writer.Write(buf[:n]); writeErr != nil {
    return fmt.Errorf("向客户端写入流式响应失败: %w", writeErr)
}

// after
if _, writeErr := c.Writer.Write(buf[:n]); writeErr != nil {
    if c.Request.Context().Err() != nil {
        return nil // client canceled
    }
    return fmt.Errorf("向客户端写入流式响应失败: %w", writeErr)
}

if _, writeErr := c.Writer.Write(buf[:n]); writeErr != nil {
    if errors.Is(writeErr, context.Canceled) || c.Request.Context().Err() != nil {
        return nil // client canceled: not an error
    }
    return fmt.Errorf("向客户端写入流式响应失败: %w", writeErr)
}

When it happens

Trigger: While LLMAuto is proxying tokens, the browser/tab is closed, the user navigates away or cancels the fetch, a proxy/load-balancer between client and Gin times out the idle/exceeding-duration connection, or the client socket errors mid-stream.

Common situations: Long LLM generations exceed an nginx/envoy proxy_read_timeout; frontend AbortController cancels the request; corporate proxies kill long-lived SSE connections; mobile clients drop networks mid-answer.

Related errors

• LLM stream request timed out
• 上游大模型流式服务返回非 2xx: status=%d content-type=%s read-body-err=%w
• 读取上游流式响应失败: %w
• 上游大模型流式服务返回非 2xx: status=%d content-type=%s read-body-err=%w
• 读取上游流式响应失败: %w

AI-assisted analysis of flipped-aurora/gin-vue-admin@3136500ef3 (2026-08-31). Data as JSON: /api/errors/ad6f56a6476bb784. Report an issue: GitHub.