Troubleshoot Fluent Tests on GitHub Runner

[ansMHanmer]

🔍 Before submitting the issue

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

🐞 Description of the bug

Issue

Fluent tests are failing on Windows on our new GitHub runner VMs being used for SAF projects. I have installed Fluent v251,v252 using the automated installer. The test output on those is a little unhelpful:

FAILED tests/e2e/test_fluent.py::TestFluent2DDPSolver::test_fluent_solver[EndToEndSolution-PIM-Desktop-252] - ansys.saf.glow._core.client_exceptions.InternalSolutionException: The solution encountered an internal error and was unable to complete the request.
!!!!!!!!!!!!!!!!!!!!!!!!!! stopping after 1 failures !!!!!!!!!!!!!!!!!!!!!!!!!!

Upon looking a little closer and just trying to do something simpler locally:

python -m pip install ansys-fluent-core
python -c "import ansys.fluent.core as pyfluent; pyfluent.logger.enable(); pyfluent.logger.set_global_level('DEBUG');solver_session = pyfluent.launch_fluent(); print(solver_session)"

It works locally after dealing with some firewall issues. The weird thing now is setting up this exact same script in CI fails:

name: Test New Windows Runners - Fluent

env:
  ANSYSLMD_LICENSE_FILE: ${{ format('1055@{0}', secrets.LICENSE_SERVER) }}

jobs:
  test-win-runners-fluent:
    name: "Test Windows Runners w/ Fluent"
    runs-on: ['self-hosted', '<redacted>']
    steps:
      - name: Install 7-Zip
        run: |
          $7zipPath = "C:\Program Files\7-Zip\7z.exe"
          if (-not (Test-Path $7zipPath)) {
            $7zipUrl = "https://www.7-zip.org/a/7z2408-x64.exe"
            $7zipInstaller = "$env:TEMP\7z-installer.exe"
            Invoke-WebRequest -Uri $7zipUrl -OutFile $7zipInstaller -UseBasicParsing
            Start-Process -FilePath $7zipInstaller -ArgumentList "/S" -Wait
          }
          echo "C:\Program Files\7-Zip" >> $env:GITHUB_PATH
        shell: powershell

      - uses: actions/setup-python@v6
        id: python
        with:
          python-version: '3.12'

      - run: |
          python -m venv .venv
          .venv\Scripts\Activate.ps1
          python -m pip install --upgrade pip
          python -m pip install ansys-fluent-core
          python -c "import ansys.fluent.core as pyfluent; pyfluent.logger.enable(); pyfluent.logger.set_global_level('DEBUG');solver_session = pyfluent.launch_fluent(); print(solver_session)"
        shell: powershell

Result:

pyfluent.launcher ERROR: Exception caught - RuntimeError: WSAGetOverlappedResult: Connection reset (An existing connection was forcibly closed by the remote host.

 -- 10054)
Traceback (most recent call last):
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\launcher\standalone_launcher.py", line 275, in __call__
    session = self.new_session._create_from_server_info_file(
              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\session.py", line 365, in _create_from_server_info_file
    fluent_connection = FluentConnection(
                        ^^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\fluent_connection.py", line 560, in __init__
    self._health_check.check_health()
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\services\health_check.py", line 82, in check_health
    response = self._stub.Check(
               ^^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\grpc\_interceptor.py", line 276, in __call__
    response, ignored_call = self._with_call(
                             ^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\grpc\_interceptor.py", line 328, in _with_call
    call = self._interceptor.intercept_unary_unary(
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\services\interceptors.py", line 174, in intercept_unary_unary
    return self._intercept_call(continuation, client_call_details, request)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\services\interceptors.py", line 164, in _intercept_call
    raise new_ex from None
RuntimeError: WSAGetOverlappedResult: Connection reset (An existing connection was forcibly closed by the remote host.

 -- 10054)

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "<string>", line 1, in <module>
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\utils\deprecate.py", line 136, in wrapper
    return deprecated_func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\deprecated\classic.py", line 199, in wrapper_function
    return wrapped_(*args_, **kwargs_)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\utils\deprecate.py", line 136, in wrapper
    return deprecated_func(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\deprecated\classic.py", line 199, in wrapper_function
    return wrapped_(*args_, **kwargs_)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\launcher\launcher.py", line 410, in launch_fluent
    return launcher()
           ^^^^^^^^^^
  File "C:\xxx\.venv\Lib\site-packages\ansys\fluent\core\launcher\standalone_launcher.py", line 319, in __call__
    raise LaunchFluentError(self._launch_cmd) from ex
ansys.fluent.core.launcher.error_handler.LaunchFluentError: 
Fluent Launch command: "C:\Program Files\ANSYS Inc\v252\fluent\ntbin\win64\fluent.exe" 3ddp -py -hidden -sifile=C:\Users\xxx\AppData\Local\Temp\serverinfo-28ng5cnh.txt -nm
PyFluent logging file C:\xxx\pyfluent.log
Setting PyFluent global logging level to DEBUG.
Setting PyFluent global logging level to DEBUG.

Log Files

Auto-Transcript Start Time:  14:33:33, 01 Apr 2026 Coordinated Universal Time
Current Directory: "<redacted>"
Information: The server has started and is running.

              Welcome to ANSYS Fluent 2025 R2.04

              Copyright 1987-2025 ANSYS, Inc. All Rights Reserved.
              Unauthorized use, distribution or duplication is prohibited.
              This product is subject to U.S. laws governing export and re-export.
              For full Legal Notice, see documentation.

Build Time: Oct  7 2025 15:34:38 EDT  Build Id: 10205

Connected License Server List:  1055@<redacted>

*********************************************
Info: Your license enables 4-way parallel execution.
For faster simulations, please start the application with the appropriate parallel options.
*********************************************

Host spawning Node 0 on machine "<redacted>" (win64).

----------------------------------------------------------------------------------
ID    Hostname       Core  O.S.         PID    Vendor
----------------------------------------------------------------------------------
n0    xxx 1/8  Windows-x64  14152  AMD EPYC 7763 64-Core Processor
host  xxx      Windows-x64  1856   AMD EPYC 7763 64-Core Processor

MPI Option Selected: intel
----------------------------------------------------------------------------------

Cleanup script file is C:\\xxx\\cleanup-fluent-xxx-1856.bat

We can see that this log file from the CI/CD run unfortunately isn't helpful. And it is unclear why this would not work when the exact same script works when running it locally. I am fairly confident that resolving this simpler issue will lead us to the solution for the failing tests as well - but I am unsure what to do next.

📝 Steps to reproduce

In a virtual environment on the affected machine, create and activate a virtual environment and run:

python -c "import ansys.fluent.core as pyfluent; pyfluent.logger.enable(); pyfluent.logger.set_global_level('DEBUG');solver_session = pyfluent.launch_fluent(); print(solver_session)"

This will complete successfully.

However, running this exact same script in a GitHub Workflow fails.

💻 Which operating system are you using?

Windows

📀 Which ANSYS version are you using?

252

🐍 Which Python version are you using?

3.12

📦 Installed packages

ansys-api-fluent==0.3.39
ansys-api-platform-instancemanagement==1.1.3
ansys-api-tools-filetransfer==0.2.0
ansys-fluent-core==0.38.1
ansys-platform-instancemanagement==1.1.2
ansys-tools-common==0.5.0
ansys-tools-filetransfer==0.2.2
ansys-units==0.10.1
certifi==2026.2.25
charset-normalizer==3.4.6
click==8.3.1
colorama==0.4.6
defusedxml==0.7.1
Deprecated==1.3.1
docker==7.1.0
googleapis-common-protos==1.73.1
grpcio==1.80.0
grpcio-health-checking==1.80.0
grpcio-status==1.80.0
idna==3.11
importlib_metadata==9.0.0
joblib==1.5.3
nltk==3.9.4
numpy==2.4.4
pandas==2.3.3
platformdirs==4.9.4
protobuf==6.33.6
python-dateutil==2.9.0.post0
pytz==2026.1.post1
pywin32==311
PyYAML==6.0.3
regex==2026.3.32
requests==2.33.1
scooby==0.11.0
six==1.17.0
tqdm==4.67.3
typing_extensions==4.15.0
tzdata==2025.3
urllib3==2.6.3
wrapt==2.1.2
zipp==3.23.0

[prmukherj]

@ansMHanmer, could you please try to set this env var: PYFLUENT_CONTAINER_INSECURE_MODE = 1, during the CI run and check if the issues is resolved?
Thank you.

[ansMHanmer]

@prmukherj, I set that var in my latest run and got the same broken result. We are not running fluent in a container, although I am not sure that's exactly what the env var is implying. Anyway just wanted to make it clear it's just a plain windows desktop installation.

[mkundu1]

@ansMHanmer That environment variable is not needed for standalone launch mode. One possibility is that Fluent process has been crashed during launch. Could you please look for any error log files like cortexerror.log or fluent*error.log in the working directory?

[ansMHanmer]

Hi @mkundu1 - that does seem to be the case here - however this is all I've got other than the .trn file that I pasted in the issue description - is there a way I can possibly elevate the logging level in either? Neither file seems to be giving me the answer

cat .\fluent-0-error.log

Error [node 0] [time 4/8/26 18:37:38] Abnormal Exit!

[prmukherj]

@ansMHanmer, to change log level you could refer: https://fluent.docs.pyansys.com/version/stable/user_guide/log.html

[ansMHanmer]

2026-04-09 11:50:12,802 pyfluent.networking   DEBUG    Can use localhost as remoting ip
2026-04-09 11:50:12,804 pyfluent.launcher     DEBUG    Launching Fluent with command: "C:\Program Files\ANSYS Inc\v252\fluent\ntbin\win64\fluent.exe" 3ddp -py -hidden -sifile=C:\Users\ansys\AppData\Local\Temp\serverinfo-tivd52_y.txt -nm
2026-04-09 11:50:13,813 pyfluent.launcher     INFO     Waiting for Fluent to launch...
2026-04-09 11:50:13,813 pyfluent.launcher     INFO     ...59 seconds remaining
2026-04-09 11:50:14,813 pyfluent.launcher     INFO     Waiting for Fluent to launch...
2026-04-09 11:50:14,813 pyfluent.launcher     INFO     ...58 seconds remaining
2026-04-09 11:50:15,814 pyfluent.launcher     INFO     Waiting for Fluent to launch...
2026-04-09 11:50:15,814 pyfluent.launcher     INFO     ...57 seconds remaining
2026-04-09 11:50:16,815 pyfluent.launcher     INFO     Waiting for Fluent to launch...
2026-04-09 11:50:16,815 pyfluent.launcher     INFO     ...56 seconds remaining
2026-04-09 11:50:17,815 pyfluent.launcher     INFO     Waiting for Fluent to launch...
2026-04-09 11:50:17,815 pyfluent.launcher     INFO     ...55 seconds remaining
2026-04-09 11:50:18,816 pyfluent.launcher     INFO     Waiting for Fluent to launch...
2026-04-09 11:50:18,816 pyfluent.launcher     INFO     ...54 seconds remaining
2026-04-09 11:50:19,817 pyfluent.launcher     INFO     Waiting for Fluent to launch...
2026-04-09 11:50:19,817 pyfluent.launcher     INFO     ...53 seconds remaining
2026-04-09 11:50:20,818 pyfluent.launcher     INFO     Waiting for Fluent to launch...
2026-04-09 11:50:20,818 pyfluent.launcher     INFO     ...52 seconds remaining
2026-04-09 11:50:21,818 pyfluent.launcher     INFO     Waiting for Fluent to launch...
2026-04-09 11:50:21,818 pyfluent.launcher     INFO     ...51 seconds remaining
2026-04-09 11:50:22,819 pyfluent.launcher     INFO     Fluent has been successfully launched.
2026-04-09 11:50:22,819 pyfluent.networking   DEBUG    GRPC_TRACE: RPC = /grpc.health.v1.Health/Check, request = {}
2026-04-09 11:50:57,392 pyfluent.launcher     ERROR    Exception caught - RuntimeError: WSAGetOverlappedResult: Connection reset (An existing connection was forcibly closed by the remote host.

@prmukherj - I do set the logging level for pyfluent to debug, above is the result. It really just shows the same - so I was wondering if there was a way to elevate the logging level for fluent itself? Maybe it could tell us more detail about why it's crashing?

[ansMHanmer]

@prmukherj - any feedback on the above question? Thank you!

[mkundu1]

@ansMHanmer The problem here is that one of the Fluent processes is getting crashed in the GitHub runner. Can you try to launch Fluent from command-line in the CI script and check if it is successfully launched? You can check whether a similar error log file is written. You can test with the same command used by PyFluent to launch Fluent (seen in the PyFluent-side traceback above):

C:\Program Files\ANSYS Inc\v252\fluent\ntbin\win64\fluent.exe" 3ddp -py -hidden -sifile=<any .txt file> -nm

[ansMHanmer]

Okay I didn't get anything extra out of the logs on the run but I did check the Windows Event Viewer and found this crash event:

Faulting application name: cx2520.exe, version: 0.0.0.0, time stamp: 0x68e56850
Faulting module name: dx11_2723_x64.hdi, version: 27.23.0.0, time stamp: 0x66e1d820
Exception code: 0xc0000005
Fault offset: 0x000000000001054d

DETAILED:

<Event xmlns="http://schemas.microsoft.com/win/2004/08/events/event">
- <System>
  <Provider Name="Application Error" Guid="{a0e9b465-b939-57d7-b27d-95d8e925ff57}" /> 
  <EventID>1000</EventID> 
  <Version>0</Version> 
  <Level>2</Level> 
  <Task>100</Task> 
  <Opcode>0</Opcode> 
  <Keywords>0x8000000000000000</Keywords> 
  <TimeCreated SystemTime="2026-04-16T13:07:57.5861579Z" /> 
  <EventRecordID>18056</EventRecordID> 
  <Correlation /> 
  <Execution ProcessID="5416" ThreadID="6796" /> 
  <Channel>Application</Channel> 
  <Computer>xxx</Computer> 
  <Security UserID="S-1-5-21-3363889236-2951689076-2575324370-500" /> 
  </System>
- <EventData>
  <Data Name="AppName">cx2520.exe</Data> 
  <Data Name="AppVersion">0.0.0.0</Data> 
  <Data Name="AppTimeStamp">68e56850</Data> 
  <Data Name="ModuleName">dx11_2723_x64.hdi</Data> 
  <Data Name="ModuleVersion">27.23.0.0</Data> 
  <Data Name="ModuleTimeStamp">66e1d820</Data> 
  <Data Name="ExceptionCode">c0000005</Data> 
  <Data Name="FaultingOffset">000000000001054d</Data> 
  <Data Name="ProcessId">0x183c</Data> 
  <Data Name="ProcessCreationTime">0x1dccda1f4473fdb</Data> 
  <Data Name="AppPath">C:\PROGRA~1\ANSYSI~1\v252\fluent\fluent25.2.0\cortex\win64\cx2520.exe</Data> 
  <Data Name="ModulePath">C:\PROGRA~1\ANSYSI~1\v252\fluent\ntbin\win64\dx11_2723_x64.hdi</Data> 
  <Data Name="IntegratorReportId">89620849-8452-4c56-9f4a-f34f89b13d3c</Data> 
  <Data Name="PackageFullName" /> 
  <Data Name="PackageRelativeAppId" /> 
  </EventData>
  </Event>

[mkundu1]

@ansMHanmer The issue is from the graphics driver of the github runner. Could you try launching without GUI/graphics?

from ansys.fluent.core import UIMode
solver_session = pyfluent.launch_fluent(ui_mode=UIMode.NO_GUI_OR_GRAPHICS)

[ansMHanmer]

@mkundu1 - that code snippet does indeed work.

@vgelbgras noticed that when trying to launch the binary manually (without pyfluent), the flag that seemed to make the crash happen was -hidden. For example:

CRASHES

& "C:\Program Files\ANSYS Inc\v252\fluent\ntbin\win64\fluent.exe" 3ddp -hidden -i test.jou -t1 > launch_test.log 2>&1

DOES NOT CRASH

& "C:\Program Files\ANSYS Inc\v252\fluent\ntbin\win64\fluent.exe" 3ddp -g -i test.jou -t1 > launch_test.log 2>&1

This -hidden flag does seem to be passed in by default when launching through pyfluent, I guess that UIMode flag is what's turning it off?

We've also noticed that when running on a working machine that we used to use in another GitHub org that the -hidden flag is explicitly passed through on Windows, and there it works - is the expectation that it should work on Windows?

So I'm wondering now if this is some kind of issue with machine configuration?

[seanpearsonuk]

We consider -hidden to be the most generally robust option for headless workflows on Windows. Therefore we have no plans to change this default at the present time.

[ansMHanmer]

Hey @seanpearsonuk - definitely understood that there is no need or desire to change this default, that's not what I'm looking for. Rather I'm just trying to figure out why something that is expected to work, is not working on my setup - I'm still not certain of the answer.