sgl-project/sglang · error · RuntimeError

Video generation failed: {error_msg}

Error message

Video generation failed: {error_msg}

What it means

RuntimeError raised inside generate_video's polling loop when the job status endpoint reports a terminal error: the status payload contains an 'error' object whose message is propagated. It means the server accepted the job but generation itself failed server-side.

Source

Thrown at python/sglang/multimodal_gen/apps/ComfyUI_SGLDiffusion/core/server_api.py:335

                        headers=self.headers,
                        timeout=30,
                    )
                    status_response.raise_for_status()
                    status = status_response.json()

                    # Reset error counter on successful request
                    consecutive_errors = 0

                    if status.get("status") == "completed":
                        return status
                    elif status.get("status") == "failed":
                        error = status.get("error", {})
                        error_msg = (
                            error.get("message", "Unknown error")
                            if error
                            else "Unknown error"
                        )
                        raise RuntimeError(f"Video generation failed: {error_msg}")
                except requests.exceptions.ConnectionError as e:
                    # Connection errors - likely server is down
                    consecutive_errors += 1
                    if consecutive_errors >= max_consecutive_errors:
                        raise RuntimeError(
                            f"Lost connection to server after {consecutive_errors} consecutive errors. "
                            f"Server may be unavailable: {str(e)}"
                        )
                except requests.exceptions.RequestException as e:
                    # Other network errors - continue polling but track errors
                    consecutive_errors += 1
                    if consecutive_errors >= max_consecutive_errors:
                        raise RuntimeError(
                            f"Network error after {consecutive_errors} consecutive failures: {str(e)}"
                        )

                time.sleep(poll_interval)

View on GitHub (pinned to 0132848349)

Solutions

1. Read the embedded error_msg — it comes from the server and names the real cause.
2. Fix the reported cause: lower resolution/seconds, free GPU memory, or correct input images.
3. Check the SGLang Diffusion server logs for the corresponding stack trace.
4. Retry with simpler parameters to confirm the pipeline works.

# validate video params against model limits before submit
assert width % 16 == 0 and height % 16 == 0, 'dimensions must be multiples of 16'

try:
    job = api.generate_video(prompt=p)
except RuntimeError as e:
    # embedded error_msg names the server-side cause; fix params accordingly
    handle_generation_error(str(e))

When it happens

Trigger: The polled job status returns status=error with an error object — e.g. OOM on the GPU, invalid size/fps/seconds parameters, unsupported video model, or a corrupted input frame image.

Common situations: Requesting resolutions the video model can't handle; VRAM exhaustion with other jobs running; passing reference images in a format the server rejects.

Related errors

• Lost connection to server after {consecutive_errors} consecu
• Network error after {consecutive_errors} consecutive failure
• Video generation timed out after {max_wait_time} seconds
• Failed to generate video: {str(e)}
• Failed to generate MiniMax-H3 video: {str(e)}

AI-assisted analysis of sgl-project/sglang@0132848349 (2026-08-28). Data as JSON: /api/errors/53d7256eaa606faf. Report an issue: GitHub.