cfdengine/pvbatch repository overview

⁠Paraview (pvbatch)

This Docker image provides the non-interactive, script-runner part of the ParaView toolkit — pvbatch.

This build is intended to be used on hardware that has no access to graphics

Note: This was supposed to be an automated build with access to the Dockerfile etc but the build time is too long - I'll add a link to the Dockerfile shortly

⁠Usage

This container provides pvbatch as an entrypoint and assumes that it will be run as a drop-in replacement for a pvbatch executable on the host system.

We use it like this:

docker run --rm -v $PWD:/data -w /data cfdengine/pvbatch YOUR_PYTHON_SCRIPT.py

Where:

• --rm removes the container on completion
• -v $PWD:/data mounts the current directory into the container at /data
• -w /data sets the working directory in the container to /data ($PWD on the host)
• cfdengine/pvbatch is this container
• YOUR_PYTHON_SCRIPT.py is the python script of ParaView commands to execute (perhaps generated by a trace of an interactive session).

⁠Build Notes

⁠Alpine

This version is built from an Alpine Linux and has been kept as small as possible to help with repeated deployment on remote computing.

⁠Entrypoint

The container sets the pvbatch command as an entry point. To override this, to start an interactive shell you could use:

docker run --rm -ti --entrypoint /bin/sh cfdengine/pvbatch

⁠tini

It includes the tini package to provide a basic init system for zombie reaping - if that sounds exciting then we'll let Phusion explain⁠ why you might need an init system in a Docker container.

⁠MPI

MPI was not enabled for this build

⁠OSMesa

Paraview is compiled against OSMesa libraries to provide offscreen (software) rendering where there is no access to the GPU. OSMesa is compiled using the Gallium llvmpipe version to be as fast as possible (for a software renderer) and to allow the SurfaceLIC plugin to be used.