calesthio/OpenMontage · error · ComfyUIError

Execution error: {msgs}

Error message

Execution error: {msgs}

What it means

ComfyUIError raised by _history_entry when the history entry for prompt_id exists and its status.status_str equals 'error'. This means execution genuinely started and a node failed server-side (model load error, OOM, bad input tensor, custom node exception). The status.messages list is embedded verbatim and typically contains an execution_error entry with node type, exception message, and traceback.

Source

Thrown at tools/_comfyui/client.py:234

            f"(local/custom workflows on modest GPUs routinely exceed the "
            f"client wait) — it was not cancelled. Poll "
            f"GET {{server_url}}/history/{prompt_id} directly, or call "
            f"generate()/execute() again with a longer timeout and this "
            f"prompt_id to resume waiting without resubmitting.",
            prompt_id=prompt_id,
        )

    def _history_entry(self, prompt_id: str) -> dict | None:
        """Return a completed history entry, or ``None`` while it is absent."""
        resp = requests.get(f"{self.server_url}/history/{prompt_id}", timeout=10)
        resp.raise_for_status()
        entry = resp.json().get(prompt_id)
        if entry is None:
            return None
        status = entry.get("status", {})
        if status.get("status_str") == "error":
            msgs = status.get("messages", [])
            raise ComfyUIError(f"Execution error: {msgs}", prompt_id=prompt_id)
        return entry

    def _history_entry_if_reachable(self, prompt_id: str) -> dict | None:
        """Best-effort history probe while the websocket remains usable."""
        try:
            return self._history_entry(prompt_id)
        except ComfyUIError:
            raise
        except Exception:
            return None

    def wait_ws(
        self,
        prompt_id: str,
        *,
        timeout: int = 600,
        interval: int = 5,
        on_progress: Callable[[dict], None] | None = None,

View on GitHub (pinned to 95e1c3d0ab)

Solutions

1. Read the embedded messages — the execution_error entry names the failing node type and exception text
2. For CUDA OOM: lower resolution/batch, use tiled VAE, fp16/lowvram variants, or a smaller checkpoint
3. For missing models: verify files under models/ match the names the workflow references
4. For custom-node crashes: update or pin the node, check ComfyUI server console for the full traceback

Example fix

# before
workflow patched with batch_size=8, width=1536, height=1536  # OOM on 8GB GPU

# after
patches = {"3": {"batch_size": 1, "width": 1024, "height": 1024}}
workflow = ComfyUIClient.patch_workflow(workflow, patches)

import requests
# before submitting heavy jobs, sanity-check VRAM and model availability
info = requests.get(f"{server_url}/object_info", timeout=10).json()
assert "CheckpointLoaderSimple" in info, "no checkpoint loader available"

try:
    entry = client.poll(prompt_id)
except ComfyUIError as e:
    if "Execution error" in str(e):
        msgs = e.args[0] if e.args else str(e)
        # locate execution_error entry: node type + exception message
        raise SystemExit(f"comfyui node failed: {msgs}")
    raise

When it happens

Trigger: poll() or any history fetch for a prompt whose workflow failed mid-run: CUDA out of memory, missing model file resolved at execution time, a custom node raising Python exceptions, invalid image dimensions for a sampler.

Common situations: VRAM exhaustion with oversized batches/resolutions on consumer GPUs; model files moved after workflow validation; custom node bugs or version incompatibilities with the installed ComfyUI; fp16 instability on some GPUs.

Related errors

• Prompt {prompt_id} did not complete within {timeout}s. The j
• Execution error: {data}
• Node errors: {json.dumps(data['node_errors'])}
• Prompt error: {json.dumps(data['error'])}
• No prompt_id in response: {data}

AI-assisted analysis of calesthio/OpenMontage@95e1c3d0ab (2026-08-15). Data as JSON: /api/errors/8cef5e68cfb0a228. Report an issue: GitHub.