What causes "重试 %d 次后仍然失败" in fish2018/pansou?

Increase the 15s context timeout in searchImpl to accommodate total backoff + slow responses.

fish2018/pansou · error

重试次后仍然失败

Error message

重试 %d 次后仍然失败

What it means

Companion to error 37: when the retry loop in ash's doRequestWithRetry ends with lastErr == nil — which happens when the loop broke early because req.Context().Err() != nil (context cancelled or deadline exceeded) — the function still must return an error, so it returns '重试 %d 次后仍然失败' without a wrapped cause.

Solutions

Increase the 15s context timeout in searchImpl to accommodate total backoff + slow responses.
Check req.Context().Err() and include it in the error so the cause is not lost.
Reduce per-attempt cost (fewer header round-trips, HTTP/2 keep-alive) so the deadline suffices.
Distinguish cancellation from timeout in callers to skip pointless re-wrapping.

Example fix

// before
return nil, fmt.Errorf("重试 %d 次后仍然失败", maxRetries)
// after
if ctxErr := req.Context().Err(); ctxErr != nil {
    return nil, fmt.Errorf("重试 %d 次后仍然失败: %w", maxRetries, ctxErr)
}
return nil, fmt.Errorf("重试 %d 次后仍然失败", maxRetries)

Defensive patterns

Strategy: retry

Validate before calling

// ensure budget fits: totalBackoff + attempts*perReqTimeout < ctxDeadline
if maxRetries*retryDelay >= 15*time.Second { reduce retries or raise deadline }

Try / catch

if err != nil && errors.Is(err, context.DeadlineExceeded) {
    // treat as timeout: increase deadline or reduce work per attempt
    return partialResults, err
}

Prevention

Size context deadlines to exceed worst-case retry totals
Propagate cancellation intent so timeout vs cancel is distinguishable
Reduce backoff ceiling when contexts are short
Include ctx.Err() in the final error for diagnosability

When it happens

Trigger: The 15-second context from searchImpl expires (or is cancelled) while attempts are still failing/ongoing, so the loop breaks via the ctx.Err() check with no lastErr recorded — ash.go:307.

Common situations: Slow upstream site that cannot answer within 15s under current backoff; caller cancelled the search; environment with high latency to the target region.

Understand the failure class

Background: Request timed out: what client-side request timeouts mean across libraries (Request timed out, TIMED_OUT, APITimeoutError) — this error's family across 39 libraries.

Related errors

AI-assisted analysis of fish2018/pansou@beaa561337 (2026-09-07). Data as JSON: /api/errors/f68747450384b450. Report an issue: GitHub.

Appendix: source

Thrown at plugin/ash/ash.go:307

		
		// 清理响应
		if resp != nil {
			resp.Body.Close()
		}
		
		lastErr = err
		
		// 如果是上下文取消或超时，不再重试
		if req.Context().Err() != nil {
			break
		}
	}
	
	if lastErr != nil {
		return nil, fmt.Errorf("重试 %d 次后仍然失败: %w", maxRetries, lastErr)
	}
	
	return nil, fmt.Errorf("重试 %d 次后仍然失败", maxRetries)
}

View on GitHub (pinned to beaa561337)

重试 次后仍然失败