dpf.server.connect_to_server does not respect timeout

[janvonrickenbach]

Before submitting the issue

• I have checked for Compatibility issues
• I have searched among the existing issues
• I am using a Python virtual environment

Description of the bug

if dpf.server.connect_to_server is called before the dpf server is ready, the timeout is not respected. Instead only a fixed number of tries (3) happens and then the connection fails.

The problem is in the function
self._check_first_call(num_connection_tryouts)

in server_types.py

Steps To Reproduce

Try to connect to a server that is not yet up => it will fail before the timeout.

Which Operating System are you using?

Linux

Which Ansys version are you using?

Ansys 2023 R1

Which Python version are you using?

3.7

Installed packages

aiohttp 3.8.3
aiosignal 1.3.1
alabaster 0.7.12
ansys-dpf-composites 0.1.dev0 D:\ANSYSDev\pydpf-composites
ansys-dpf-core 0.6.2
ansys-dpf-gate 0.2.3
ansys-dpf-gatebin 0.2.3
ansys-grpc-dpf 0.6.1
ansys-sphinx-theme 0.4.2
appdirs 1.4.4
async-timeout 4.0.2
attrs 22.1.0
Babel 2.11.0
beautifulsoup4 4.11.1
bleach 5.0.1
build 0.8.0
cachetools 5.2.0
certifi 2022.9.24
cfgv 3.3.1
charset-normalizer 2.1.1
colorama 0.4.6
commonmark 0.9.1
coverage 6.5.0
cycler 0.11.0
distlib 0.3.6
docutils 0.19
exceptiongroup 1.0.4
filelock 3.8.0
fonttools 4.38.0
frozenlist 1.3.3
google-api-core 2.10.2
google-api-python-client 2.66.0
google-auth 2.14.1
google-auth-httplib2 0.1.0
googleapis-common-protos 1.57.0
grpcio 1.50.0
httplib2 0.21.0
identify 2.5.9
idna 3.4
imageio 2.22.4
imagesize 1.4.1
importlib-metadata 4.13.0
iniconfig 1.1.1
jaraco.classes 3.2.3
Jinja2 3.1.2
keyring 23.11.0
kiwisolver 1.4.4
MarkupSafe 2.1.1
matplotlib 3.5.3
more-itertools 9.0.0
multidict 6.0.2
mypy 0.931
mypy-extensions 0.4.3
nodeenv 1.7.0
numpy 1.23.5
numpydoc 1.5.0
packaging 21.3
pep517 0.13.0
Pillow 9.3.0
pip 22.2.2
pkginfo 1.8.3
platformdirs 2.5.4
pluggy 1.0.0
pre-commit 2.20.0
protobuf 4.21.9
psutil 5.9.4
pyasn1 0.4.8
pyasn1-modules 0.2.8
pydata-sphinx-theme 0.9.0
Pygments 2.13.0
pypandoc 1.10
pyparsing 3.0.9
pytest 7.2.0
pytest-cov 3.0.0
python-dateutil 2.8.2
pytz 2022.6
pyvista 0.36.1
pywin32-ctypes 0.2.0
PyYAML 6.0
readme-renderer 37.3
requests 2.28.1
requests-toolbelt 0.10.1
rfc3986 2.0.0
rich 12.6.0
rsa 4.9
scooby 0.7.0
setuptools 65.6.0
setuptools-scm 6.4.2
six 1.16.0
snowballstemmer 2.2.0
soupsieve 2.3.2.post1
Sphinx 5.3.0
sphinx_autodoc_typehints 1.19.5
sphinx-copybutton 0.4.0
sphinx-gallery 0.11.1
sphinxcontrib-applehelp 1.0.2
sphinxcontrib-devhelp 1.0.2
sphinxcontrib-htmlhelp 2.0.0
sphinxcontrib-jsmath 1.0.1
sphinxcontrib-qthelp 1.0.3
sphinxcontrib-serializinghtml 1.1.5
toml 0.10.2
tomli 2.0.1
tqdm 4.64.1
twine 4.0.1
typing_extensions 4.4.0
uritemplate 4.1.1
urllib3 1.26.12
virtualenv 20.16.7
vtk 9.2.2
webencodings 0.5.1
wheel 0.37.1
wslink 1.9.1
yarl 1.8.1
zipp 3.10.0

[laurent-ge]

@rafacanton @PProfizi
Yes, in my attempt, it fails after 3.3s when I had defined a timeout of 10s.

DPFServerException                        Traceback (most recent call last)
Cell In[2], line 1
----> 1 dpf.connect_to_server(as_global=False, timeout=10)

File [c:\ANSYSDev\dpf_py_venv\lib\site-packages\ansys\dpf\core\server.py:382](file:///C:/ANSYSDev/dpf_py_venv/lib/site-packages/ansys/dpf/core/server.py:382), in connect_to_server(ip, port, as_global, timeout, config, context)
    380 server_type = ServerFactory.get_remote_server_type_from_config(config)
    381 try:
--> 382     return connect()
    383 except ModuleNotFoundError as e:
    384     if "use a LegacyGrpcServer" in e.msg:

File [c:\ANSYSDev\dpf_py_venv\lib\site-packages\ansys\dpf\core\server.py:364](file:///C:/ANSYSDev/dpf_py_venv/lib/site-packages/ansys/dpf/core/server.py:364), in connect_to_server..connect()
    359 server_init_signature = inspect.signature(server_type.__init__)
    360 if (
    361     "ip" in server_init_signature.parameters.keys()
    362     and "port" in server_init_signature.parameters.keys()
    363 ):
--> 364     server = server_type(
    365         ip=ip,
    366         port=port,
    367         as_global=as_global,
    368         launch_server=False,
    369         context=context,
    370     )
    371 else:
    372     server = server_type(as_global=as_global, context=context)

File [c:\ANSYSDev\dpf_py_venv\lib\site-packages\ansys\dpf\core\server_types.py:821](file:///C:/ANSYSDev/dpf_py_venv/lib/site-packages/ansys/dpf/core/server_types.py:821), in GrpcServer.__init__(self, ansys_path, ip, port, timeout, as_global, load_operators, launch_server, docker_config, use_pypim, num_connection_tryouts, context)
    819 self.live = True
    820 self._create_shutdown_funcs()
--> 821 self._check_first_call(num_connection_tryouts)
    822 if context:
    823     if context == core.AvailableServerContexts.no_context:

File [c:\ANSYSDev\dpf_py_venv\lib\site-packages\ansys\dpf\core\server_types.py:843](file:///C:/ANSYSDev/dpf_py_venv/lib/site-packages/ansys/dpf/core/server_types.py:843), in GrpcServer._check_first_call(self, num_connection_tryouts)
    839 except errors.DPFServerException as e:
    840     if ("GOAWAY" not in str(e.args) and "unavailable" not in str(e.args)) or i == (
    841         num_connection_tryouts - 1
    842     ):
--> 843         raise e

File [c:\ANSYSDev\dpf_py_venv\lib\site-packages\ansys\dpf\core\server_types.py:837](file:///C:/ANSYSDev/dpf_py_venv/lib/site-packages/ansys/dpf/core/server_types.py:837), in GrpcServer._check_first_call(self, num_connection_tryouts)
    835 for i in range(num_connection_tryouts):
    836     try:
--> 837         self.version
    838         break
    839     except errors.DPFServerException as e:

File [c:\ANSYSDev\dpf_py_venv\lib\site-packages\ansys\dpf\core\server_types.py:860](file:///C:/ANSYSDev/dpf_py_venv/lib/site-packages/ansys/dpf/core/server_types.py:860), in GrpcServer.version(self)
    858     major = integral_types.MutableInt32()
    859     minor = integral_types.MutableInt32()
--> 860     api.data_processing_get_server_version_on_client(self.client, major, minor)
    861     self._version = str(int(major)) + "." + str(int(minor))
    862 return self._version

File [c:\ANSYSDev\dpf_py_venv\lib\site-packages\ansys\dpf\gate\generated\data_processing_capi.py:498](file:///C:/ANSYSDev/dpf_py_venv/lib/site-packages/ansys/dpf/gate/generated/data_processing_capi.py:498), in DataProcessingCAPI.data_processing_get_server_version_on_client(client, major, minor)
    496 res = capi.dll.DataProcessing_getServerVersion_on_client(client._internal_obj if client is not None else None, ctypes.byref(utils.to_int32(major)), ctypes.byref(utils.to_int32(minor)), ctypes.byref(utils.to_int32(errorSize)), ctypes.byref(sError))
    497 if errorSize.value != 0:
--> 498 	raise errors.DPFServerException(sError.value)
    499 return res

DPFServerException: a 'service is currently unavailable' error occurred: failed to connect to all addresses

[PProfizi]

The timeout argument is currently used only when starting the server.
Another function then checks the connection made by trying to run self.version. This function currently only tries 3 times (default number of time) to get the server to give its version. There is no notion of time.
The parameter for the number of times the client tries to get the server version is currently not even exposed through connect_to_server.

It may be interesting to switch from a logic of number of tries to a logic of timeout, probably reusing the timeout for server start.

[PProfizi]

@laurent-ge can you please try again with this branch?
https://github.com/ansys/pydpf-core/tree/fix/connect_to_server_does_not_respect_timeout

[PProfizi]

We should try to come up with a test for this. I am not sure how to start a server but fail to ask its version. We may need to create mock-up servers in the testing framework to test this kind of things.
FYI @rafacanton

[laurent-ge]

@laurent-ge can you please try again with this branch? https://github.com/ansys/pydpf-core/tree/fix/connect_to_server_does_not_respect_timeout

@PProfizi , yes I just tested and it works well, thanks.

Tested:

• run connect_to_server when no server is available, check that timeout value is respected
• on a python console, run connect_to_server(timeout=60). On another python console, wait 30s and run start_local_server(as_global=True). Check in the first console that the server gets connected (grpc).

[rafacanton]

@laurent-ge @PProfizi great! Let's merge and close!