-
Notifications
You must be signed in to change notification settings - Fork 0
Overview
Once the bbp image has been built, it can be run with this script: $ run_bbp.sh
This script mounts the ./target subdirectory in the container. It will pull the bbp image specified in this file from a local Docker repo, or from Dockerhub if not found locally. The images are 50GB+, so if they are downloaded from Dockerhub, it may take more than an hour to start up the container. Once the image is cached locally, the container will start up rapidly.
The user runs the run_bbp.sh script in terminal window at the command line on their host computer. When they run this script, the command line changes in their terminal window, and the user is presented with a Linux command line prompt. The users should cd into the /app/target directory. The files from their host computer (e.g. the files in /Users/maechlin/bbp_docker/target) will found in this directory. When the users runs the bbp platform from here, the results will be written to subdirectories of /app/target. $cd /app/target
Users can invoke the BBP command line interface, and the bbp will ask users a series of questions about the simulation they wish to run. $run_bbp.py
docker run --rm -it --mount type=bind,source="$(pwd)"/target,destination=/app/target sceccode/bbp_docker:MMDDHHMM
This is a coding and configuration test for creating a UCVM docker image that can be run on AWS.
There is a .dockerignore file that defines which files not to include in the image. The Dockerfile and this README.md are excluded.
The top level script is: build_all.sh which invokes docker build 1 time.
This lists the steps needed to build the container. It starts with a amazonlinux base image, add compilers and python.
It copies the ucvm git repo from the build computer into the image, and then invokes the build process. The build process runs, installs results in a directory: /app/bbp
On host system, user invokes docker run. Expectation is that there is a subdirectory call ./target ./target is mounted as /app/bbp_data. Input files can be stored there. Output results will be written there
Potential Limitations
Users must be comfortable running ucvm from a command line interface. This over means they are creating output files, and extracting selected information for plotting. Users must work within limits of images and local computers. There are some size ucvm problems that won't run on their laptops, so we need to warn people what the limits are.