Fastapi: Issue: Fatal Python error: Cannot recover from stack overflow.

Created on 13 Aug 2020  路  15Comments  路  Source: tiangolo/fastapi

First check

  • [ ] I added a very descriptive title to this issue.
  • [x] I used the GitHub search to find a similar issue and didn't find it.
  • [x] I searched the FastAPI documentation, with the integrated search.
  • [x] I already searched in Google "How to X in FastAPI" and didn't find any information.
  • [x] I already read and followed all the tutorials in the docs and didn't find an answer.
  • [ ] I already checked if it is not related to FastAPI but to Pydantic.
  • [ ] I already checked if it is not related to FastAPI but to Swagger UI.
  • [ ] I already checked if it is not related to FastAPI but to ReDoc.
  • [ ] After submitting this, I commit to one of:

    • Read open issues with questions until I find 2 issues where I can help someone and add a comment to help there.

    • I already hit the "watch" button in this repository to receive notifications and I commit to help at least 2 people that ask

Description

I have this issue on my QA server, but I was not able to replicate it locally.
My code based on the latest version of https://github.com/tiangolo/full-stack-fastapi-postgresql using the base class in CRUD,
this error happened on the update function

Fatal Python error: Cannot recover from stack overflow.

Thread 0x00007f615d0c0700 (most recent call first):
  File "/usr/local/lib/python3.7/concurrent/futures/thread.py", line 78 in _worker
  File "/usr/local/lib/python3.7/threading.py", line 870 in run
  File "/usr/local/lib/python3.7/threading.py", line 926 in _bootstrap_inner
  File "/usr/local/lib/python3.7/threading.py", line 890 in _bootstrap

Thread 0x00007f615d8c1700 (most recent call first):
  File "/usr/local/lib/python3.7/concurrent/futures/thread.py", line 78 in _worker
  File "/usr/local/lib/python3.7/threading.py", line 870 in run
  File "/usr/local/lib/python3.7/threading.py", line 926 in _bootstrap_inner
  File "/usr/local/lib/python3.7/threading.py", line 890 in _bootstrap

Current thread 0x00007f615e19f700 (most recent call first):
  File "<string>", line 1 in <lambda>
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/operators.py", line 432 in __getitem__

Environment

  • OS: [e.g. Linux / Windows / macOS]: ubuntu
  • FastAPI Version [e.g. 0.3.0]: 0.61.0
  • Python version: 3.7

Additional context

  def update(self, db: Session, *, db_obj: ModelType, obj_in: Union[UpdateSchemaType, Dict[str, Any]]) -> ModelType:
    obj_data = jsonable_encoder(db_obj)
    if isinstance(obj_in, dict):
      update_data = obj_in
    else:
      update_data = obj_in.dict(exclude_unset=True)
    for field in obj_data:
      if field in update_data:
        setattr(db_obj, field, update_data[field])
    db.add(db_obj)
    db.commit()
    db.refresh(db_obj)
    return db_obj
question
Source
picture imadmoussa1

Most helpful comment

This was caused by the new release (0.61.0) (jsonable_encoder).

The problem is the line that contains: obj_data = jsonable_encoder(db_obj)

Possible solution:

# Your function
  def update(self, db: Session, *, db_obj: ModelType, obj_in: Union[UpdateSchemaType, Dict[str, Any]]) -> ModelType:
    obj_data = db_obj.as_dict() # HERE!
    if isinstance(obj_in, dict):
      update_data = obj_in
    else:
      update_data = obj_in.dict(exclude_unset=True)
    for field in obj_data:
      if field in update_data:
        setattr(db_obj, field, update_data[field])
    db.add(db_obj)
    db.commit()
    db.refresh(db_obj)
    return db_obj

# SQLAlchemy Base object
@as_declarative()
class Base:
    def as_dict(self) -> dict:
        return {c.key: getattr(self, c.key) for c in inspect(self).mapper.column_attrs}

But I don't know why we have this error message, I need to check.

EDIT: As reminded by @imadmoussa1:

from sqlalchemy import inspect
picture Kludex on 13 Aug 2020
👍5

All 15 comments

This was caused by the new release (0.61.0) (jsonable_encoder).

The problem is the line that contains: obj_data = jsonable_encoder(db_obj)

Possible solution:

# Your function
  def update(self, db: Session, *, db_obj: ModelType, obj_in: Union[UpdateSchemaType, Dict[str, Any]]) -> ModelType:
    obj_data = db_obj.as_dict() # HERE!
    if isinstance(obj_in, dict):
      update_data = obj_in
    else:
      update_data = obj_in.dict(exclude_unset=True)
    for field in obj_data:
      if field in update_data:
        setattr(db_obj, field, update_data[field])
    db.add(db_obj)
    db.commit()
    db.refresh(db_obj)
    return db_obj

# SQLAlchemy Base object
@as_declarative()
class Base:
    def as_dict(self) -> dict:
        return {c.key: getattr(self, c.key) for c in inspect(self).mapper.column_attrs}

But I don't know why we have this error message, I need to check.

EDIT: As reminded by @imadmoussa1:

from sqlalchemy import inspect
Kludex picture Kludex on 13 Aug 2020
👍5

I have the same FATAL errors when running normal select statement with databases/encode.

async def list_creators(
    skip: int = 0,
    limit: int = 0,
    manager_id: Optional[int] = None,
    mapping_group_id: Optional[int] = None,
):
    # basic select statement
    base_query = models.Creator.__table__.select()
    # if input manager_id
    # add where condition of mapping_groups with manager_id
    if manager_id:
        base_query = base_query.where(
            models.Creator.mapping_group_id.in_(
                select([column("mapping_group_id")])
                .select_from(models.MappingGroupMember.__table__)
                .where(models.MappingGroupMember.manager_id == manager_id)
            )
        )
    # if mapping_group_id
    # add where condition of mapping_groups with mapping_group_id
    if mapping_group_id:
        base_query = base_query.where(
            models.Creator.mapping_group_id == mapping_group_id
        )
    # add skip and limit
    query = base_query.offset(skip).limit(limit) if limit else base_query.offset(skip)
    return await database.fetch_all(query)

@router.get(
    "/list",
)
async def read_creators(
    skip: int = 0,
    limit: int = 100,
    manager_id: Optional[int] = None,
    mapping_group_id: Optional[int] = None,
):
    if manager_id and mapping_group_id:
        raise HTTPException(
            status_code=400,
            detail="Please do not provide manager_id and mapping_group_id at the same time",
        )
    return await crud.list_creators(
        skip=skip, limit=limit, manager_id=manager_id, mapping_group_id=mapping_group_id
    )

Any help would be appreciated. This did not happen before.

anh-collab picture anh-collab on 13 Aug 2020

Same issue here, this crash does not give any trace info, took some time to trace the offending code, as outlined earlier:
https://github.com/tiangolo/full-stack-fastapi-postgresql/blob/master/%7B%7Bcookiecutter.project_slug%7D%7D/backend/app/app/crud/base.py#L49

duz-sg picture duz-sg on 14 Aug 2020

@Kludex Thank you, Your solution fixed my problem.

imadmoussa1 picture imadmoussa1 on 14 Aug 2020
🚀1

You're welcome! @imadmoussa1 Do you mind closing the issue then? :sunglasses: :tada:

Kludex picture Kludex on 14 Aug 2020
👍1 
def as_dict(self) -> dict:
        return {c.key: getattr(self, c.key) for c in inspect(self).mapper.column_attrs}

Hi @Kludex
I applied the fix, but it throws me below error:

  File ".\crud\base.py", line 49, in update
    obj_data = db_obj.as_dict()
  File ".\db\base_class.py", line 15, in as_dict
    return {c.key: getattr(self, c.key) for c in inspect(self).mapper.column_attrs}
TypeError: 'module' object is not callable
sv8083 picture sv8083 on 15 Aug 2020

Hi @sv8083 you should import the inspect from sqlalchemy

from sqlalchemy import inspect

imadmoussa1 picture imadmoussa1 on 15 Aug 2020
👍2

Does it mean jsonable_encoder is broken? It is unclear why the issue was closed.

plankter picture plankter on 17 Aug 2020
👍3

@plankter jsonable_encoder is not broken, I am using it in different functions, but on update function we had errors. fixed by replacing the jsonable_encoder with a custom function.

imadmoussa1 picture imadmoussa1 on 18 Aug 2020

This fix does not help in my case, after applying the above fix, I still get the error message:

Fatal Python error: Cannot recover from stack overflow.

Thread 0x00007fa28e7bc700 (most recent call first):
  File "/usr/local/lib/python3.7/concurrent/futures/thread.py", line 78 in _worker
  File "/usr/local/lib/python3.7/threading.py", line 870 in run
  File "/usr/local/lib/python3.7/threading.py", line 926 in _bootstrap_inner
  File "/usr/local/lib/python3.7/threading.py", line 890 in _bootstrap

Thread 0x00007fa28effd700 (most recent call first):
  File "/usr/local/lib/python3.7/concurrent/futures/thread.py", line 78 in _worker
  File "/usr/local/lib/python3.7/threading.py", line 870 in run
  File "/usr/local/lib/python3.7/threading.py", line 926 in _bootstrap_inner
  File "/usr/local/lib/python3.7/threading.py", line 890 in _bootstrap

Current thread 0x00007fa28f7fe700 (most recent call first):
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/type_api.py", line 67 in operate
  File "<string>", line 1 in <lambda>
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/operators.py", line 432 in __getitem__
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/elements.py", line 762 in operate

Update, fixed my issue:

    check_item = crud.item.get_by_user(
        db=db, id=item_id, user_id=current_user.id
    )
    pd_check_item = ItemCheck.from_orm(check_item) # Fix: convert from DB model to pydantic model 
    check_data = jsonable_encoder(pd_check_item, exclude_none=True) # Breaks if accept db model directly
    return check_data

It seems jsonable_encoder can no longer accept a DB model and convert the data, it only accepts a pydantic model.

duz-sg picture duz-sg on 18 Aug 2020
👍2

Another approach I was trying locally was to use load a pydantic model from the orm model, get the dict(), then iterate that, such as:

obj_data = UserInDB.from_orm(db_obj).dict()

...now I know that doesnt perfectly fit into CRUDBase unless we refactor it to also hold onto a DBSchemaType in addition to CreateSchemaType and UpdateSchemaType

shawnwall picture shawnwall on 18 Aug 2020
👍1

This fix does not help in my case, after applying the above fix, I still get the error message:

Fatal Python error: Cannot recover from stack overflow.

Thread 0x00007fa28e7bc700 (most recent call first):
  File "/usr/local/lib/python3.7/concurrent/futures/thread.py", line 78 in _worker
  File "/usr/local/lib/python3.7/threading.py", line 870 in run
  File "/usr/local/lib/python3.7/threading.py", line 926 in _bootstrap_inner
  File "/usr/local/lib/python3.7/threading.py", line 890 in _bootstrap

Thread 0x00007fa28effd700 (most recent call first):
  File "/usr/local/lib/python3.7/concurrent/futures/thread.py", line 78 in _worker
  File "/usr/local/lib/python3.7/threading.py", line 870 in run
  File "/usr/local/lib/python3.7/threading.py", line 926 in _bootstrap_inner
  File "/usr/local/lib/python3.7/threading.py", line 890 in _bootstrap

Current thread 0x00007fa28f7fe700 (most recent call first):
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/type_api.py", line 67 in operate
  File "<string>", line 1 in <lambda>
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/operators.py", line 432 in __getitem__
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/elements.py", line 762 in operate

Update, fixed my issue:

    check_item = crud.item.get_by_user(
        db=db, id=item_id, user_id=current_user.id
    )
    pd_check_item = ItemCheck.from_orm(check_item) # Fix: convert from DB model to pydantic model 
    check_data = jsonable_encoder(pd_check_item, exclude_none=True) # Breaks if accept db model directly
    return check_data

It seems jsonable_encoder can no longer accept a DB model and convert the data, it only accepts a pydantic model.

This fix does not help in my case, after applying the above fix, I still get the error message:

Fatal Python error: Cannot recover from stack overflow.

Thread 0x00007fa28e7bc700 (most recent call first):
  File "/usr/local/lib/python3.7/concurrent/futures/thread.py", line 78 in _worker
  File "/usr/local/lib/python3.7/threading.py", line 870 in run
  File "/usr/local/lib/python3.7/threading.py", line 926 in _bootstrap_inner
  File "/usr/local/lib/python3.7/threading.py", line 890 in _bootstrap

Thread 0x00007fa28effd700 (most recent call first):
  File "/usr/local/lib/python3.7/concurrent/futures/thread.py", line 78 in _worker
  File "/usr/local/lib/python3.7/threading.py", line 870 in run
  File "/usr/local/lib/python3.7/threading.py", line 926 in _bootstrap_inner
  File "/usr/local/lib/python3.7/threading.py", line 890 in _bootstrap

Current thread 0x00007fa28f7fe700 (most recent call first):
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/type_api.py", line 67 in operate
  File "<string>", line 1 in <lambda>
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/operators.py", line 432 in __getitem__
  File "/usr/local/lib/python3.7/site-packages/sqlalchemy/sql/elements.py", line 762 in operate

Update, fixed my issue:

    check_item = crud.item.get_by_user(
        db=db, id=item_id, user_id=current_user.id
    )
    pd_check_item = ItemCheck.from_orm(check_item) # Fix: convert from DB model to pydantic model 
    check_data = jsonable_encoder(pd_check_item, exclude_none=True) # Breaks if accept db model directly
    return check_data

It seems jsonable_encoder can no longer accept a DB model and convert the data, it only accepts a pydantic model.

How to import ItemCheck

l1725821492 picture l1725821492 on 20 Aug 2020

How to import ItemCheck

I'm using the project generator, so you can define your own pydantic models like ItemCheck, you can refer to the samples in project generator here:
https://github.com/tiangolo/full-stack-fastapi-postgresql/blob/master/%7B%7Bcookiecutter.project_slug%7D%7D/backend/app/app/schemas/item.py#L33
As long as it is based on ItemInDBBase, it will support orm_mode

duz-sg picture duz-sg on 20 Aug 2020
👍1

For me, and also using sqlalchemy, I was simply forgetting to add response_model=schemaName to my endpoint:

Before:

@app.get("/users/me")
async def read_users_me(current_user: schemas.User = Depends(get_current_active_user)):
    return current_user

After:

@app.get("/users/me", response_model=schemas.User)
async def read_users_me(current_user: schemas.User = Depends(get_current_active_user)):
    return current_user
chrisjsimpson picture chrisjsimpson on 23 Aug 2020
👍1

Thanks for the discussion here everyone! I'm sorry for the inconveniences :facepalm: . This was fixed in 0.61.1. The current latest is 0.62.0.

tiangolo picture tiangolo on 9 Dec 2020
Was this page helpful?
0 / 5 - 0 ratings

Related issues

levchik picture levchik  路  3Comments
RogerioDosSantos picture RogerioDosSantos  路  3Comments
PyDataBlog picture PyDataBlog  路  3Comments
cjw296 picture cjw296  路  3Comments
zero0nee picture zero0nee  路  3Comments