Containerized version of PFLOTRAN
If you have a dedicated directory for a model to hold the input and output
files, first go to that directory in your shell. For example,
if you have a directory called model_files in your home directory, then
issue:
$ cd ~/model_files
Suppose you have a model input file in that directory called exampleinfile.in
You can invoke PFLOTRAN with:
$ docker run -v $(pwd):/data subsurfaceinsights/pflotran -pflotranin ./exampleinfile.in
It then run and produce output files in the directory
You can run a model in parallel by setting the NUM_PROCESSES environment
variable. Following the previous example:
$ docker run -e NUM_PROCESSES=6 -v $(pwd):/data subsurfaceinsights/pflotran -pflotranin ./exampleinfile.in
PFLOTRAN will then be invoked with mpirun -np 6 pflotran to run with 6
parallel processes
The container's working directory is /data so mapping any directory to that
data will provide access to model files.
The container's PFLOTRAN process will run as whatever user owns the /data
directory to maintain ownership consistency.