youngkingdom/dify
1
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-06-08 09:27:39 +08:00
Code Issues Packages Projects Releases Wiki Activity

fix: keep cleanup tasks resilient to billing API failures (#35600)

Browse Source
This commit is contained in:
zyssyz123
2026-04-27 16:51:09 +08:00
committed by fatelei
parent e22b03797c
commit e73f720505
6 changed files with 479 additions and 31 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
api/tasks/clean_document_task.py
View File

@ -61,13 +61,31 @@ def clean_document_task(document_id: str, dataset_id: str, doc_form: str, file_i
# check segment is exist
if index_node_ids:
index_processor = IndexProcessorFactory(doc_form).init_index_processor()
with session_factory.create_session() as session:
dataset = session.scalar(select(Dataset).where(Dataset.id == dataset_id).limit(1))
if dataset:
index_processor.clean(
dataset, index_node_ids, with_keywords=True, delete_child_chunks=True, delete_summaries=True
)
# Wrap vector / keyword index cleanup in try/except so that a transient
# failure here (e.g. billing API hiccup propagated via FeatureService when
# ModelManager is initialized inside ``Vector(dataset)``) does not abort
# the entire task and leave document_segments / child_chunks / image_files
# / metadata bindings stranded in PG. Mirrors the pattern already used in
# ``clean_dataset_task`` so the document row's hard delete (already
# committed by the caller) does not produce orphan PG rows just because
# the vector backend or one of its transitive dependencies was unhappy.
try:
index_processor = IndexProcessorFactory(doc_form).init_index_processor()
with session_factory.create_session() as session:
dataset = session.scalar(select(Dataset).where(Dataset.id == dataset_id).limit(1))
if dataset:
index_processor.clean(
dataset, index_node_ids, with_keywords=True, delete_child_chunks=True, delete_summaries=True
)
except Exception:
logger.exception(
"Failed to clean vector / keyword index in clean_document_task, "
"document_id=%s, dataset_id=%s, index_node_ids_count=%d. "
"Continuing with PG / storage cleanup; vector orphans can be reaped later.",
document_id,
dataset_id,
len(index_node_ids),
)
total_image_files = []
with session_factory.create_session() as session, session.begin():