Regression-tests failing inside a container

[Wei-Keng Liao]

I set these environment variables at configure time, for example,

    ./configure --prefix=$HOME/PnetCDF \
                --enable-profiling \
                --with-mpi=$MPI_PATH \
                TESTSEQRUN="$MPI_PATH/bin/mpirun -n 1” \
                TESTMPIRUN="$MPI_PATH/bin/mpiexec -n NP” \
                TESTOUTDIR=“/scratch”

so, I do not need to set them when running make.


Wei-keng

> On Mar 19, 2020, at 3:26 PM, Carl Ponder <cponder at nvidia.com> wrote:
> 
> On 3/19/20 1:25 PM, Wei-Keng Liao wrote:
>> Can you try this?
>>     make -i -k check TESTSEQRUN="mpirun -n 1”
>> or
>>     make -i -k check TESTSEQRUN=“srun -n 1”
>> 
>> i.e. setting the environment variables in the make command line.
>> 
> Ok that nailed it. Any idea why I have to do this? Since I'm installing all the Ubuntu packages into the container image, I'm wondering if I'm missing something that the configure script is looking for.
> I'm seeing similar failures with some of the NetCDF regression-tests and trying to figure out if there is an equivalent workaround.
> Thanks,
> 
>                    Carl
> 
> This email message is for the sole use of the intended recipient(s) and may contain confidential information.  Any unauthorized review, use, disclosure or distribution is prohibited.  If you are not the intended recipient, please contact the sender by reply email and destroy all copies of the original message.