iflytek/astron-agent · error · CustomException

ChunkDeleteFailed

ChunkDeleteFailed

Error message

Deletion failed: {error_msg}

What it means

chunks_delete reached RAGFlow successfully but RAGFlow returned a non-success response for the delete operation. The strategy extracts response['message'] (default 'Deletion failed') and raises CustomException with ChunkDeleteFailed prefixed by 'Deletion failed:'. This is an application-level failure reported by the RAGFlow API, not a transport exception.

Solutions

1. Inspect the message after 'Deletion failed:' — it is RAGFlow's own error text and usually names the offending chunk or reason.
2. Verify docId and chunkIds still exist via query_doc before deleting; refresh stale IDs.
3. Confirm the dataset_id used to build the RAGFlow request matches the document's actual dataset.
4. Treat 'not found' style messages as idempotent success if your workflow allows concurrent deletes.

Example fix

// before
await strategy.chunks_delete(docId, chunk_ids)
// after
try:
    await strategy.chunks_delete(docId, chunk_ids)
except CustomException as e:
    if "not found" in str(e).lower():
        logger.warning("Chunks already gone, treating as success")
    else:
        raise

existing = await strategy.query_doc(docId)
existing_ids = {c["id"] for c in existing}
deletable = [cid for cid in chunk_ids if cid in existing_ids]

null

try:
    await strategy.chunks_delete(docId, chunk_ids)
except CustomException as e:
    if e.code == CodeEnum.ChunkDeleteFailed and "not found" in str(e).lower():
        logger.warning("Chunks already deleted: %s", e)
    else:
        raise

When it happens

Trigger: RAGFlow delete-chunk API returns code != 0 — e.g. chunk already deleted, chunk ID not found in the dataset/document, wrong dataset_id/docId pairing, or RAGFlow-side permission/validation rejection.

Common situations: Stale chunk IDs in the UI after someone else deleted the chunk; document recreated so chunk IDs no longer match; RAGFlow dataset permissions changed.

Related errors

• REPO_FILE_DELETE_FAILED
• REPO_KNOWLEDGE_DELETE_FAILED
• fetch_all_document_chunks failed on page
• RAGFLOW_RAGError
• OPERATION_FAILED

AI-assisted analysis of iflytek/astron-agent@5e758547a8 (2026-09-12). Data as JSON: /api/errors/2b4443a7a136bf32. Report an issue: GitHub.

Appendix: source

Thrown at core/knowledge/service/impl/ragflow_strategy.py:1059

            logger.info(f"Using dataset: {dataset_id}")

            # 2. Call RAGFlow deletion API directly
            delete_response = await ragflow_client.delete_chunks(
                dataset_id=dataset_id, document_id=docId, chunk_ids=chunkIds
            )

            logger.info(f"RAGFlow chunk deletion response: {delete_response}")

            # 3. Process response
            if delete_response.get("code") == 0:
                logger.info(f"Successfully deleted {len(chunkIds)} chunks")
                return None  # Success, let API layer handle the response
            else:
                # RAGFlow deletion failed
                error_msg = delete_response.get("message", "Deletion failed")
                logger.error(f"RAGFlow deletion failed: {error_msg}")
                raise CustomException(
                    CodeEnum.ChunkDeleteFailed, f"Deletion failed: {error_msg}"
                )

        except CustomException:
            raise  # Re-raise custom exceptions
        except Exception as e:
            logger.error(f"Chunk deletion operation failed: {e}")
            raise CustomException(
                CodeEnum.ChunkDeleteFailed, f"Deletion operation failed: {str(e)}"
            )

    async def query_doc(self, docId: str, **kwargs: Any) -> List[Dict[str, Any]]:
        """Query all chunk information for a document using RAGFlow."""
        try:
            logger.info(f"Starting document chunk query: docId={docId}")

            dataset_id = await self._resolve_dataset_id(kwargs.get(_DATASET_ID_KWARG))
            if not dataset_id:

View on GitHub (pinned to 5e758547a8)