Published references, source identities and connected reading.
Have the exact error text? Paste it into Troubleshoot to find matching articles →
Log keyword: execution_cached
A fast completion can involve cached execution, a preview-only endpoint, temporary storage or a client misreading events. It is not automatically a failed generation.
Log keyword: Address already in use
Repeatedly starting ComfyUI can collide with an existing healthy instance. A port conflict is not permission to terminate every Python process.
Log keyword: UnetLoaderGGUF
GGUF is a container format, not a guarantee that an arbitrary chat-model file can serve as a diffusion model or text encoder in a particular ComfyUI node.
Log keyword: No module named yaml
Match the import name to its distribution before changing the Python environment that runs ComfyUI.
Log keyword: No module named aiohttp
Check library availability before investigating DNS, TLS, HTTP responses or timeouts.
Log keyword: Failed to get custom node list
Distinguish registry retrieval, search filters, installed code and Python imports.
Log keyword: Failed to save workflow
Preserve an export before investigating draft storage, file downloads or API-format confusion.
Log keyword: Failed to fetch
Locate the failing endpoint and distinguish the backend, logs, registry and workflow-save requests.
Log keyword: mat1 and mat2 must have the same dtype
A dtype failure differs from a shape or device failure. Start with the traceback and incompatible operand types; a KSampler error alone does not identify the component that caused them.
Log keyword: Failed to fetch dynamically imported module
The Python server may still be running while the canvas fails. Distinguish static assets, JavaScript extensions and an absent backend before reinstalling unrelated models.
Log keyword: invalid value encountered in cast
Compare the saved artifact with the preview before diagnosing a black image. A black preview and genuinely invalid output pixels are different observations.
Log keyword: prompt_outputs_failed_validation
A rejected request needs structured error inspection. HTTP 400 is not, by itself, evidence of an unstable network and should not trigger endless resubmission.
Log keyword: Unknown encoder
Existing output can still be unplayable, silent or the wrong length. Establish whether the issue belongs to generation, encoding, saving or the target player.
Log keyword: HeaderTooLarge
A .safetensors header parsing failure can involve an incomplete or wrong download, a different format or a parser/loader mismatch. The message alone does not uniquely prove corruption.
Log keyword: The paging file is too small for this operation to complete
This Windows error concerns committed system memory. It should not be treated as VRAM exhaustion or proof that a safetensors file is corrupt merely because it appears during model loading.
Log keyword: torch.cuda.is_available() = False
For an intended NVIDIA workload, False only says that this process cannot use CUDA. Check the build, device visibility and driver rather than treating it as a complete diagnosis.
Log keyword: 401 Client Error
Seeing a model card does not guarantee that a script is authorized to fetch weights. Verify resource access and credential identity rather than bypassing TLS or repository restrictions.
Log keyword: ControlNet
Diagnose ControlNet compatibility, control-image preparation and conditioning separately; node-link type errors need their own check.
Log keyword: execution_error
A connected socket and accepted queue request are not a result. Check client_id, prompt_id, terminal events and history together.
Log keyword: CLIP input is invalid: None
A wire reaching CLIPTextEncode does not prove that a valid encoder was loaded. Check upstream loading and whether the model uses a complete checkpoint or separate components.
Log keyword: lora key not loaded
A selectable LoRA can still target an incompatible model, or the workflow can bypass its modified outputs. Little visible change is an observation, not proof of a damaged file.
Log keyword: No space left on device
A target output disk can have free space while a download cache, temporary directory, container writable layer or another mounted filesystem is full.
Log keyword: MPS backend out of memory
On Apple Silicon, applications, browsers, model weights and tensors share memory resources. Total installed memory is not all available to one generation task.
Log keyword: No module named 'triton'
A missing Python module, a missing Triton dependency and a Windows _fused DLL failure occur at different layers. Upgrading every dependency together is not a targeted repair.