Learning Forum

[v.bhatnagar.1]

Hello everyone,

I have been trying to run a .wbpj file on my University's Linux-based HPC using Ansys 2024 R1, but unable to do so due to an error. Even when I submit an empty "minimal_test.wbpj" file, it returns the following error. The error is the same regardless of whether the .wbpj file is my actual simulation setup (on Ansys Polyflow) or just an empty file for testing purposes.

I am attaching all relevant files for you to check. I use a SLURM script to run it on my University's HPC cluster.

Error file:

_____

_____

Job started at: Fri 23 May 12:39:25 BST 2025

SLURM_JOB_ID: 7752

Attempting to run Workbench...

 

Framework error caught :

 

Fatal error: An unexpected error has occurred: Unable to load project '/home/postgrads/2913170B/DieSwell_Simple/minimal_test.wbpj': Ansys.Core.Events.EventPublishFailedException: Failed to create object 6a7e2939-5ffe-47a0-bda7-fb034a0e3a0f, error code hres: 8000FFFF ---> Ansys.Core.Events.EventPublishFailedException: Failed to create object 6a7e2939-5ffe-47a0-bda7-fb034a0e3a0f, error code hres: 8000FFFF ---> Ansys.Core.Events.EventPublishFailedException: Failed to create object 6a7e2939-5ffe-47a0-bda7-fb034a0e3a0f, error code hres: 8000FFFF ---> Ansys.Core.Events.EventPublishFailedException: Failed to create object 6a7e2939-5ffe-47a0-bda7-fb034a0e3a0f, error code hres: 8000FFFF ---> System.Runtime.InteropServices.COMException: Failed to create object 6a7e2939-5ffe-47a0-bda7-fb034a0e3a0f, error code hres: 8000FFFF

  at Ans.ComponentSystem.Interop.MiniCom.AnsCoCreateInstance (System.Type classType, System.IntPtr aggr) [0x0007f] in <44fa2ffd3745466592ae7a1a4b470645>:0 

  at Ansys.ACT.Interop.DefinitionIdentifier.ClassFactoryCreateObjectDelegate (System.IntPtr aggr) [0x00000] in <90c9343f97d04f79bbc2eb7d1584c740>:0 

  at System.__ComObject.CreateIUnknown (System.Type t) [0x00015] in :0 

  at Mono.Interop.ComInteropProxy.CreateProxy (System.Type t) [0x00000] in :0 

  at System.Runtime.Remoting.RemotingServices.CreateClientProxyForComInterop (System.Type type) [0x00000] in :0 

  at System.Runtime.Remoting.Activation.ActivationServices.CreateProxyForType (System.Type type) [0x0003b] in :0 

  at (wrapper managed-to-native) System.Object.__icall_wrapper_ves_icall_object_new_specific(intptr)

  at Ansys.ACT.WorkBench.EventObserver.Receive (Ansys.Core.Events.IEvent eventInfo) [0x0029d] in :0 

  at Ansys.Core.Events.EventObserverProxy.Receive (Ansys.Core.Events.IEvent eventInfo) [0x00000] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.EventInvoker.Invoke () [0x001a8] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

   --- End of inner exception stack trace ---

  at Ansys.Core.Events.EventInvoker.Invoke () [0x00213] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.EventPublishHelper+<>c.b__8_0 (System.Object obj) [0x00000] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.ThreadPoolWait.HandleWorkItem (System.Object state) [0x00007] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

   --- End of inner exception stack trace ---

  at Ansys.Core.Events.ThreadPoolWait.RethrowExceptions () [0x00051] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.ThreadPoolWait.WaitOne (System.Int32 millisecondsTimeout, System.Boolean exitContext) [0x00076] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.ThreadPoolWait.WaitOne () [0x00000] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.EventPublishHelper.FireAsyncWait (System.Collections.Generic.List1[T] events) [0x0006d] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.EventManager.FireEvent (System.Collections.Generic.List1[T] observers, Ansys.Core.Events.IEvent eventInfo, System.Boolean wait) [0x00064] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.EventManager.fireEventsByPriority (System.Collections.Generic.List1[T] observers, Ansys.Core.Events.IEvent eventInfo, System.Boolean wait, Ansys.Core.Events.EventObserverPriority priority) [0x00065] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.EventManager.PublishEvent (Ansys.Core.Events.IEvent eventInfo, System.Collections.Generic.List1[T] matchingProxies, System.Boolean wait) [0x00053] in <2269a6352d194a2cb2dedbf711bef8d1>:0
...

at Ansys.Core.Events.EventManager.PublishEvent (Ansys.Core.Events.IEvent ev) [0x00008] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.EventManager.DoPublishAndWait (Ansys.Core.Events.IEvent eventInformation) [0x00098] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Events.EventManager.PublishAndWait (Ansys.Core.Events.IEvent eventInformation) [0x00078] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Persistence.Commands.OpenCommand.LoadProjectInternal (Ansys.Core.Commands.IFullContext context, System.String projectFilePath, System.Boolean mergedProject, System.Boolean importProject, Ansys.Core.DataModel.ProjectSystem.SharedAccess+AccessLock shareLock, System.Boolean skipStartingEvents, System.Collections.Generic.IReadOnlyDictionary`2[TKey,TValue] addinsToPostMigrate) [0x000d2] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

CommandName: Open(FilePath="minimal_test.wbpj")

  at Ansys.Core.Commands.CommandAsyncResult.RethrowFailureException (System.Exception e) [0x0000c] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Commands.CommandAsyncResult.Wait (System.Int32 milliSecondsTimeout, System.Boolean exitContext) [0x00095] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Commands.CommandAsyncResult.Wait () [0x00000] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ansys.Core.Persistence.Commands.OpenCommand.InvokeAndWait (Ansys.Core.Commands.IProtectedContext context, System.String FilePath) [0x00023] in <2269a6352d194a2cb2dedbf711bef8d1>:0 

  at Ans.Program.Runtime.OpenFile () [0x00022] in :0 

  at Ans.Program.Runtime.DoRun () [0x0014d] in :0

 

Ansys Workbench batch process (Update and Save) completed at: Fri 23 May 12:39:38 BST 2025

_____

_____

 

SLURM script: Modified using AI, since I am new to Linux and batch runs.

_____

_____

#!/bin/bash

 

#SBATCH --account=YOUR_ACTUAL_ACCOUNT # <-- VERIFY OR REMOVE

#SBATCH --time=00-03:00                # Time (DD-HH:MM)

#SBATCH --mem=16G                      # Total Memory

#SBATCH --ntasks=4                     # Number of cores

#SBATCH --nodes=1                      # Do not change

#SBATCH --partition=standard           # Use the correct partition

 

echo "Job started at: $(date)"

echo "SLURM_JOB_ID: $SLURM_JOB_ID"

# ... (other echo statements for SLURM vars if you want)

 

module load ansys/2024-R1   # Corrected module name

 

# Remove lock file - this is good

rm -fv *_files/.lock

 

# The following lines for modifying global SolveHandlers.xml are DANGEROUS and will likely fail.

# It's better to configure parallelism within the Workbench project GUI.

# MWFILE=/software/apps/ansys_inc/2024-R1/v241/aisol/DesignSpace/DSPages/Language/en-us/xml/SolveHandlers.xml

# if [ "$SLURM_NNODES" == 1 ]; then

#   MEMPAR=0

# else

#   MEMPAR=1

# fi

# sed -re "s/(.AnsysSolution>+)[a-zA-Z0-9]*(<\/Distribute.)/\1$MEMPAR\2/" -i "$MWFILE"

# sed -re "s/(.Processors>+)[a-zA-Z0-9]*(<\/MaxNumber.)/\1$SLURM_NTASKS\2/" -i "$MWFILE"

# sed -i "s!UserConfigured=\"0\"!UserConfigured=\"1\"!g" "$MWFILE"

 

export KMP_AFFINITY=disabled

export I_MPI_HYDRA_BOOTSTRAP=ssh

# export ENABLE_ANSYS_QA_SERVICES241=0 # Kept from our previous script, likely no harm

 

echo "Attempting to run Workbench..."

/software/apps/ansys_inc/2024-R1/v241/Framework/bin/Linux64/runwb2 -B -E "Update();Save(Overwrite=True)" -F minimal_test.wbpj

 

echo "Ansys Workbench batch process (Update and Save) completed at: $(date)"

_____

_____

The original SLURM script format (provided by the University) is as follows:

#!/bin/bash

#SBATCH --account=def-account
#SBATCH --time=00-03:00                # Time (DD-HH:MM)
#SBATCH --mem=16G                      # Total Memory (set to 0 for all node memory)
#SBATCH --ntasks=4                     # Number of cores
#SBATCH --nodes=1                      # Do not change (multi-node not supported)
##SBATCH --exclusive                   # Uncomment for scaling testing
##SBATCH --constraint=broadwell        # Applicable to graham or cedar

module load ansys/2024-R1   # OR newer Ansys modules

if [ "$SLURM_NNODES" == 1 ]; then
  MEMPAR=0                               # Set to 0 for SMP (shared memory parallel)
else
  MEMPAR=1                               # Set to 1 for DMP (distributed memory parallel)
fi

rm -fv *_files/.lock
MWFILE=/software/apps/ansys_inc/2024-R1/v241/aisol/DesignSpace/DSPages/Language/en-us/xml/SolveHandlers.xml
sed -re "s/(.AnsysSolution>+)[a-zA-Z0-9]*(<\/Distribute.)/\1$MEMPAR\2/" -i "$MWFILE"
sed -re "s/(.Processors>+)[a-zA-Z0-9]*(<\/MaxNumber.)/\1$SLURM_NTASKS\2/" -i "$MWFILE"
sed -i "s!UserConfigured=\"0\"!UserConfigured=\"1\"!g" "$MWFILE"

export KMP_AFFINITY=disabled
export I_MPI_HYDRA_BOOTSTRAP=ssh

/software/apps/ansys_inc/2024-R1/v241/Framework/bin/Linux64/runwb2 -B -E "Update();Save(Overwrite=True)" -F DieSwell_Simple_2025-05-19.wbpj

_____

_____

Any suggestions on fixing this error are welcome. Please feel free to ask any follow-up questions; I will respond.

[Rob]

The .wbpj file is just a header for a Workbench workflow. Did you archive up the project and move the archive to HPC to then extract and run? 

[v.bhatnagar.1]

Thanks for the reply, Rob. If I understood you correctly, I did transfer the whole folder containing the simulation files to the HPC, and not just the .wbpj file.

This error occurs with the whole folder copied to the HPC.

[Rob]

Did you transfer the file & folder or archive and transfer that? The latter ensures that all paths are set to the location where it's unarchived. 

[v.bhatnagar.1]

I don't know what "archive" is, so I probably didn't do this. I copied and pasted the folder containing the project files.

[Rob]

That may have lost the links. Under Workbench>File you'll see Archive, create one and then move that to the HPC. 

[v.bhatnagar.1]

Thanks a lot. So, what would the procedure look like? Workbench - File - Archive - Copy-Paste to HPC. How do I extract the archived file on the HPC? And, subsequently, is there any change in the way I run it using SLURM?

[Rob]

Not sure about SLURM, that should call whatever you need to run. 

[v.bhatnagar.1]

Okay, I'll figure out the SLURM bit. But I am not entirely sure what to do once I archive the project on Workbench. Please explain this bit to me.

[Rob]

You copy the archive over & extract for use as before. Workbench stores paths in the .wbpj file from the original system and these may not be correct when you move to the HPC.

[v.bhatnagar.1]

Thanks a lot. I'll try this and get back to you within the next 2 days to confirm if it's working or not.

[Rob]

You're welcome. If it's still not working I'll flag for the Install specialists - I just drive the software. 

[v.bhatnagar.1]

Update:

I spoke to an Ansys representative, who suggested the following:

_____

Assuming you are running Polyflow – here is a detailed explanation. Please review this and let me know if this answers your questions and addresses the issue or not.

Firstly lets us break Simulation / Anslysis into 2 parts

1. 1. Pre and post – this is where you might import geometry, cleanup, build mesh and boundary conditions and generate an input file or files for the solver to use. And later analyze results, plot graphs and images etc. this part is interactive meaning you will usually use a Graphical User Interface (GUI) and will involve lot of keyboard and mouse inputs and feedback on a graphics terminal (monitor).
       
      Also in this phase you will typically work on a workbench project which includes the .wbpj project definition and a folder names progjectname_files which contain various files that are needed.
       
      The pre and post work is done on workstations / laptops / desktops.
      Most HPC resources tend to be batch only (no GUI) and typically do not run
       
   2. The other part is Solution cell / solve where the solver runs, reads in input files and generates output files.
      the input files typically are not the whole workbench project and folder.

Typically – Workbench is not run on HPC as a batch job, although that is an option. Workbench requires many other graphics related dependencies which may not necessarily be available in the batch (commandline – not GUI) HPC environments.

For Polyflow here is what you can try:

First please review the Overview of the Folders for a Polyflow Classic-Based Project in Workbench at this link

https://ansyshelp.ansys.com/public/account/secured?returnurl=//Views/Secured/corp/v251/en/poly_pfwb/pwb_wb_sec_wb_file_struct.html

You will need the .msh (or .poly) file and .dat file (.dat has the name of the mesh file)

Copy those to a separate folder if necessary on a file system which is accessible across the HPC

Then you can run a job using slurm using a script like example below. Please modify or add to this script to include any other slurm specific information. Your IT administrator can help with linux paths, queue or partition names etc, if you are not familiar with those yet.

#!/bin/bash
##SBATCH --account="Username_here"
#SBATCH --mail-user="email@company.com_here"
#SBATCH --mail-type=ALL
#SBATCH --job-name="SUP"
#SBATCH --error=%x.out
#SBATCH --partition=partition_name_here
#SBATCH --time=24:00:00
##
## Environment variables setup
## Polyflow Command
/apps/ansys_inc/v241/polyflow/bin/polyflow < polyflow.dat > listing.lst -th 8

The -th 8 is for number of cores, you will also need to specify that on your sbatch command

Once you run a case on cluster for polyflow, then you copy the entire post folder contents to local drive. Then use cfd-post to read the .res file to post-process the results in local workstation.

_____

Additionally, when I tried this on the HPC, there was an error: "error while loading shared libraries: libnsl.so.1: cannot open shared object file: No such file or directory". This error was fixed by requesting the University's HPC team to install the "libnsl" library on the cluster. After this, the simulation runs on the HPC. Hope this comes in handy!