[arindam]

Hi

I am trying to configure Slum in a Ubuntu Cluster. After installing Slurm and configuring, when I try to run script below, I get an error.

#!/bin/bash

# Command line args: case cells iterations


# The name of the script is myjob
#SBATCH --export=ALL
#SBATCH -J fluent
#SBATCH --partition=normal
#SBATCH --account=xrg
#SBATCH --time=01:00:00
# Number of nodes
#SBATCH -N 1
# Number of MPI processes per node
#SBATCH --ntasks-per-node=4
# The Journal file
JOURNALFILE=test.journal

# Total number of Processors
# NPROCS=4
NTASKS=`echo $SLURM_TASKS_PER_NODE | cut -c1-2`
NPROCS=`expr $SLURM_NNODES \* $NTASKS`
fluent 3ddp -g -slurm -t4 -mpi=intel -path/opt/ansys_lic/v222/fluent -ssh -i $JOURNALFILE > fluent-out.log


This is the error I receive.

/opt/ansys_inc/v222/fluent/fluent22.2.0/cortex/lnamd64/cortex.22.2.0: error while loading shared libraries: libGL.so.1: cannot open shared object file: No such file or directory

I am able to launch fluent normally with and without GUI if I use X2Go or Putty. So, fluent is installed currently. Right now, I am running fluent using non graphics option from fluent, but wanted to have the Slurm working for productivity.

Did anyone know how to resolve this error?



Thanks

Arindam

[kepler123]

I think the fix is quite simple. I have had the issue with another software, not fluent. But when you are missing a library file you simply need to add a PATH field to your bash file.

You probably have it set-up on your bashrc file but when you run with slurm you are not using your own user account anymore so the bashrc is not recognised. This is how it is for me, at least, on the cluster I use.

[arindam]

Thanks for your help. I have located that library in /usr/lib/x86_64-linux-gnu.

dpkg -L libgl1
/.
/usr
/usr/lib
/usr/lib/x86_64-linux-gnu
/usr/lib/x86_64-linux-gnu/libGL.so.1.7.0
/usr/share
/usr/share/bug
/usr/share/bug/libgl1
/usr/share/bug/libgl1/control
/usr/share/doc
/usr/share/doc/libgl1
/usr/share/doc/libgl1/copyright
/usr/lib/x86_64-linux-gnu/libGL.so.1
/usr/share/doc/libgl1/changelog.Debian.gz

and then modified my script

#!/bin/bash

# Command line args: case cells iterations


# The name of the script is myjob
#SBATCH --export=ALL
#SBATCH -J fluent
#SBATCH --partition=normal
#SBATCH --account=xrg
#SBATCH --time=01:00:00


# Number of nodes
#SBATCH -N 1

# Number of MPI processes per node
#SBATCH --ntasks-per-node=4

export FLUENT_GUI=off
FLUENT_ARCH=lnamd64 export FLUENT_ARCH
export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu:$LD_LIBRARY_PATH
# The Journal file
JOURNALFILE=test.journal



# Total number of Processors
# NPROCS=4
NTASKS=`echo $SLURM_TASKS_PER_NODE | cut -c1-2`
NPROCS=`expr $SLURM_NNODES \* $NTASKS`


fluent 3ddp -g -slurm -t4 -mpi=intel -path/opt/ansys_lic/v222/fluent -ssh -i $JOURNALFILE > fluent-out.log


But still it gives the same error. Can you please let me know if I am in the correct way?

[kepler123]

export LD_LIBRARY_PATH=/usr/lib/x86_64-linux-gnu:$LD_LIBRARY_PATH

What you have above is probably also fine and just a syntax difference, but I would write it as:

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/lib/x86_64-linux-gnu
echo $LD_LIBRARY_PATH

The echo just to make sure your bash file even reaches this point.