Error when running MFiX 19.3.1 and MFIX 20.1.0

MFiX
#1

Hi,

I’ve been running into an issue when trying to run a DEM simulation of a rotating drum on my university cluster (I’m getting the same issue when using the rotating drum tutorial as well). Everything compiles correctly but when I try to run I receive the following errors:

From: INIT_ERROR_MANAGER
Error 1000: RUN_NAME too short.
From: INIT_ERROR_MANAGER
Error 1001: Failed to open log file: .LOG
Aborting run.
/gpfs/projects/dufeklab/awinner3/DrUm/mfixsolver: line 3: 26904 Segmentation fault (core dumped) env PYTHONPATH="/gpfs/projects/dufeklab/awinner3/DrUm":"":${PYTHONPATH:+:$PYTHONPATH} /home/awinner3/anaconda3/envs/mfix-20.1.0/bin/python3.7 -m mfixgui.pymfix “$@”

Any thoughts?

Thanks,

Amelia

DrUm.mfx (10.5 KB)

#2

Hi Amelia -

This is very strange - the attached input file contains the line

run_name = 'DrUm'

which is a perfectly acceptable run name, but the job is acting as though it’s set to an empty string - note that the log file is just called .LOG whereas it should be called
DRUM.LOG - (the solver forces everything into upper-case, so your mixed-case DrUm becomes DRUM.)

I’m trying to run your case here in an attempt to reproduce this, but I’m missing an stl file - either geometry.stl or geometry.stl.original - can you send the rest of the files needed to run this job? (You can use the “bug report” to create a .zip file)

  • Charles
#3

Hi Charles,

Sure thing. The zip file is attached. Thank you for your help!

drum_2020-04-07T104937.160310.zip (141.0 KB)

#4

Hi Amelia.

I’m getting an error when running your case, but it’s different than the error you are seeing:

MFiX running: elapsed time 0:00:00  
 MFiX (20.1.0-dev) simulation on host: x280
 Run name: DRUM                  Time: 11: 8                    Date:  4- 9-2020
 Memory required:    9.00 Mb

INFO open_files.f:123
The following mesh file was not found: DRUM.msh
Mesh generation will be performed and a mesh file will be saved.

The following mesh file was successfully opened: DRUM.msh
 From Get_Stl_Data: Analyzing BCs...
  BC_ID    BC_TYPE
 Number of CG BCs =      0

ERROR cartesian_grid/get_stl_data.f:671
ERROR: NO CARTESIAN GRID BOUNDARY CONDITION SPECIFIED.
AT LEAST ONE BC_TYPE MUST START WITH CG (FOR EXAMPLE CG_NSW)
RUN ABORTED

Note in particular that the output says:

Run name: DRUM

as expected. The error message you attached shows an empty run name (it’s trying to open a file called .LOG because the run name is an empty string). This is a bit of a mystery to me at this time.

Can you tell me any more about the cluster you are running on, and how the MFIX solver binary was built?

#5

I can’t find the message right now, but I seem to recall a similar problem being caused by a Fortran library version mismatch. Does the issue only happen when running on your cluster? Are you able to run these simulations locally?

#6

Hi Charles,

I attempted to remove anaconda and MFIX and found some other things. When I attempt to install gcc make on my cluster I get an error:

(mfix-20.1.2) [awinner3@n224 ~]$ conda install gcc make

Collecting package metadata (current_repodata.json): done

Solving environment: failed with initial frozen solve. Retrying with flexible solve.

Collecting package metadata (repodata.json): done

Solving environment: failed with initial frozen solve. Retrying with flexible solve.

PackagesNotFoundError: The following packages are not available from current channels:

** - gcc**

Current channels:

** - https://repo.anaconda.com/pkgs/main/linux-64**

** - https://repo.anaconda.com/pkgs/main/noarch**

** - https://repo.anaconda.com/pkgs/r/linux-64**

** - https://repo.anaconda.com/pkgs/r/noarch**

To search for alternate channels that may provide the conda package you’re

looking for, navigate to

https://anaconda.org

and use the search bar at the top of the page.

I attempted to fix this issue by running:

conda install -c anaconda gcc make

but then if I try to launch MFIX, I get the following error:

(mfix-20.1.2) [awinner3@n224 ~]$ Traceback (most recent call last):

** File “/home/awinner3/anaconda3/envs/mfix-20.1.2/lib/python3.7/site-packages/qtpy/init.py”, line 204, in **

from PySide import version as PYSIDE_VERSION # analysis:ignore

ModuleNotFoundError: No module named 'PySide’

During handling of the above exception, another exception occurred:

Traceback (most recent call last):

** File “/home/awinner3/anaconda3/envs/mfix-20.1.2/bin/mfix”, line 6, in **

from mfixgui.gui import main

** File “/home/awinner3/anaconda3/envs/mfix-20.1.2/lib/python3.7/site-packages/mfixgui/gui.py”, line 27, in **

from qtpy import QtGui, QtCore, QtWidgets

** File “/home/awinner3/anaconda3/envs/mfix-20.1.2/lib/python3.7/site-packages/qtpy/init.py”, line 210, in **

raise PythonQtError(‘No Qt bindings could be found’)

qtpy.PythonQtError: No Qt bindings could be found

Not sure if the above is related to the previous issue or not. I also noticed I only get this run name error when trying to use the project solver. If I use the default solver, it doesn’t have an issue picking up the run name.

Best,

Amelia

#7

The Anaconda gcc package is somewhat old; if your system has a regular GCC installation that might work better.

As of MFiX 20.1, you can also install MFiX with Pip. Download the MFiX source tarball from the Download page. You will need Python 3.7. If your cluster’s system Python isn’t 3.7, you could use Conda for the Python itself but then use Pip for MFiX and it’s dependencies:

> conda create -n mfix-pip python=3.7
> conda activate mfix-pip
> pip install mfix-20.1.2-tar.gz
1 Like
#8

I’m running MFiX on a CC. I know it’s recommended to use MFiX using pip, but will the benefits of doing so remain if I run pip through conda?

#9

Recommend installing MFiX with conda, then if that doesn’t work try installing with pip. Using pip in a conda environment should work, but isn’t tested as much. Please let us know how it works out.

It’s usually easier when installing Python applications (like MFiX) to use separate environments (in case you encounter problems), either Python Virtual Environments or Conda environments