Error in Parallel Fluent
 

hi all,

I am trying to run Fluent v14.5 on 16 cores and am having a problem.

Sometimes when I click cancel to interrupt the iterations, it stops as normal then around a second later Fluent disappears without any warning.

When I check the error log file it says:
"Error [node 999999] [time 7/23/13 12:21:41]".

Does anyone have any idea how I can fix this?

Any help would be greatly appreciated.

Phil

sorry I left out a bit at the end of the error message, the full message should read:

"Error [node 999999] [time 7/23/13 12:21:41] Abnormal Exit!"

any help would be hugely appreciated!!!

the same problem
 

hi,
I countered the same problem ,too.Is there anybody can help me out?Thanks

Hello!

I see that you opened the thread some months ago. Did you find a solution?? Because I am facing the same problem!!!!! :(

Thank you!

Although I am not sure but this may have something to do with node-master communication.

I also suggest you try different method for mesh partition for parallel calculations.

what do you mean with node-master communication??

The problem happens with the version 14.X, when I run the case with Fluent 12.X, there is no problem...

Before closing the window automatically, it shows "[...]fluent14.5.7\cortex\win64\cx1457.exe received fatal signal (ACCESS_VIOLATION)"


I tried to execute just the cx1457.exe, and it says that a dll file is missing, but it is actually there!!!

Any clue??

Thank you!

I am not a cluster geek but all I understand is this: if you dont have your MPI set right then there could be issues with communication between nodes.

The 'node 999999' error also has something to do with UDFs. If your UDF is not compiled with right program (32 bit-64 bit) or right way then there could be some issues.

If you using academic license then it also comes with some limitations e.g. no. of node that can be used for parallel process.

I'm also having the same problem.

Any help would be appreciated.

Whats exactly is the problem with ansys?

I tried different methods still am getting the same..:(

How to overcome from "abnormal exit of fluent"??

Whats exactly you do during or after which the fluent closes abruptly?

I'm thinking that problem may be due to the boundary conditions, even I tried changing the BCs. But still it is getting the same error and disappearing without giving any information. I'm not understanding what to with this..:(


"Error node [99999] Abnormal Exit"..

Any help would be appreciated...


Thanks,
Shivakumar

I use ansys15.0 to simulate a transmission,I also encounter the same problem sa yours,have you solved it?
Thank you!

Try using a different MPI or playing with your MPI settings.

I tried everything discussed in this forum still getting the same error [abnormal exit]. I am using another system to run..

Hello;

I do not know exactly how to solve the problem you are facing: But i will share my experience of parallel fluent -

PBS File looks like:

### PBS ###
#PBS -l nodes=2 : ppn=8 ( Ok there is no space between : and ppn but I had to use it because a weird emoticon :p << keeps coming up)
#PBS -l walltime=30:00:00
#PBS -j oe
#PBS -o FILE.log
#PBS -V
#PBS -N Jobname
cd $PBS_O_WORKDIR
cd /home/usr [ This is where you are submitting your job from so figure it]
/data1/ansys/v150/fluent/bin/fluent 3ddp -i pressurebased.jou -t16 -g -mpi=openmpi -ssh >> logfile


OK So using these settings you can create the batch file. Now for the journal file make sure you have it saved in unix format: You can definitely create it in windows also, and transfer it to the linux system: Open the file using VIM editor ( that's what I use) and follow these commands:

:w ( write the file once more)
: set ff=unix ( set the file format as unix)
:wq ( write and quit the file)

hope it gets sorted now: :) if not come back and ask i will try to help

Thanks for the reply.. Nothing worked for me, still the fluent is spitting the same error.

I know this is an old thread, but did you have any luck with this problem?

This is usually an intermitten problem with parallel FLuent and it's usually a problem with the MPI. Is your problem intermittent or does it always happen? Try using a different MPI or reconfiguring your current MPI.

hi,
I tried every settings which are discussed in this forum but couldn't solve it.
Finally, I Uninstalled ANSYS 14.0 and installed 15.0 version, the problem disappeared automatically.

Thanks,
shiv

LuckyTran, thank you very much for your reply.

I checked and double checked everything in the case, but finally I solved the problem by setting up the case again from cero.

Just a suggestion to this problem that might help someone:
I solved an exact problem when I chose to set my inlet (mass-flow) direction (direction specification method) as " Normal to boundary".
Basically as I can understand it was going backwards or something.

I hope this helps.

Frank

Check your display driver
 

Hi folks,

I'm going to add in my solution. I'm using Fluent 16 and I also recently had issues with the program having random crashes (both during problem setup and during a solution run). I noticed that just before crashing, the monitor would usually flick off for a second or two before coming back with Fluent gone. I had recently changed out video cards and had Fluent using the GPU for a boost (GPGPU). I suspect there was a problem with the GPU calls being misread by the new GPU, resulting in the critical failures. Once I unchecked the box, Fluent worked like normal again.

So if you are having this random crashing problem and you are using at least Fluent 15, you might check to see if it's a GPU acceleration issue.

Edit: I checked my Windows event viewer and found that it was, indeed, the display driver that is causing the problem.

I'm less inclined to think it is the GPGPU now, as I found that the problem had more to do with running a case that involved relatively high airflow through a smallish tube. I was using the target-mass-flow-rate option when it happened. I found that turning that off and running at a lower flow rate, made things work better. It seems to have something to do with long calculation times between iterations, and some weird call problem to the display driver.

In fact, minimizing the window and letting the program run seemed to keep the error at bay. So whatever the problem is, it appears to be related to Nvidia and its drivers.

Failure
 

Can anyone help me. I have similar problem. Ansys Fluent is working for about an hour and then disappears. It is leaving behind error files equal to number of threads that he is working on, displayed in notepad files but only node 0 file has a message inside:

Error [node 0] [time 14.06.2017 12:37:21] Abnormal Exit!

and an additional file with number 999999 with same text but coresponding number.

Error
 

Please help me.

fluent error 99999
 

I have the facing same problem which is error 99999 in parallel fluent. Fluent disappears without any warning.
any advice and suggestions will be greatly appreciated Thanks.

This is also a problem I am facing. Moreover, I have been in contact with ANSYS for weeks. I have to say this is very disappointing after having invested and having worked with ANSYS all these years.

I really think ANSYS should provide all the info and support to users facing this simple yet abstract error: "Abnormal exit"!!!

The problems related to the compilation of UDF codes remain unsolved as well by the way! There is no support record nonwhatsoever in their customer portal as well!

fluent error
 

[QUOTE=shivakumar;545206]Thanks for the reply.. Nothing worked for me, still the fluent is spitting the same error.[/QUOTE

i also experienced in the fluent error 9999 in the log file and it causes the unexpected closing of fluent file and lost all data, i think its problem with licence if you are using academic version and you using parellel running it cause fluent error 9999 try to use serial running option sorry any kind of mistake from me iam new member of this online forum

Hey guys,
Am using ANSYS 2019 R1 and I am also experiencing this "Error [node 0] [time 4/22/20 0:18:37] Abnormal Exit!" error. I doubt it has something to do with the ANSYS version. Did anyone ever find a solution?

Hello

So, I had the same problem and was stuck for a while. But managed finally.
If anyone does get this error, then you can give it a try with this method.

1) Start a new Fluent file with serial processing.
2) Import your mesh file and run your simulation for about few iterations
3) Export the case file
4) Start a new Fluent file with parallel processing
5) Import the case file and initialise
6) Now you should be able to run the simulation without Fluent shutting down

Although this is not an eloquent solution, but helped get around the problem.

Regards