Merge pull request #1145 from UV-CDAT/docker

Adding Docker - Adding Dockerfiles for generating docker images

Author: mattben

.dockerignore

@@ -0,0 +1,11 @@
+.git
+Packages/DV3D/demo
+Packages/visus/ext-libs
+Packa
+Packages/demo
+Packages/cmor/Test
+contrib/ZonalMeans/Test
+Packages/vcs_legacy
+Packages/WK/Test
+contrib/HDF5Tools/Examples
+Packages/cmor/Doc

CMake/dashboard/docker.cmake

@@ -0,0 +1,21 @@
+set(CTEST_SOURCE_DIRECTORY /usr/src/uvcdat)
+set(CTEST_BINARY_DIRECTORY /tmp/uvcdat-build)
+
+include(${CTEST_SOURCE_DIRECTORY}/CTestConfig.cmake)
+set(CTEST_SITE "Docker Ubuntu:14.04")
+set(CTEST_BUILD_NAME "nogui-master")
+set(CTEST_CMAKE_GENERATOR "Unix Makefiles")
+
+ctest_start("Experimental")
+ctest_configure(OPTIONS "-DCDAT_BUILD_WEB=ON;-DCDAT_BUILD_GUI=OFF;-DCDAT_BUILD_OSMESA=ON;-DCDAT_BUILD_OFFSCREEN=ON;-DCMAKE_INSTALL_PREFIX=/opt/uvcdat")
+ctest_build()
+ctest_test(PARALLEL_LEVEL 4 RETURN_VALUE res)
+ctest_coverage()
+file(REMOVE "${CTEST_BINARY_DIRECTORY}/coverage.xml")
+ctest_submit()
+
+file(REMOVE "${CTEST_BINARY_DIRECTORY}/test_failed")
+if(NOT res EQUAL 0)
+  file(WRITE "${CTEST_BINARY_DIRECTORY}/test_failed" "error")
+  message(FATAL_ERROR "Test failures occurred.")
+endif()

Dockerfile

@@ -0,0 +1,18 @@
+FROM uvcdat/ubuntu
+MAINTAINER UV-CDAT Developers <uvcdat-support@llnl.gov>
+
+RUN mkdir -p /tmp/uvcdat-build
+ADD . /usr/src/uvcdat
+RUN cd /tmp/uvcdat-build && cmake -DCDAT_BUILD_WEB=ON -DCDAT_BUILD_GUI=OFF -DCDAT_BUILD_OSMESA=ON -DCDAT_BUILD_OFFSCREEN=ON -DCMAKE_INSTALL_PREFIX=/opt/uvcdat /usr/src/uvcdat && make && cd / && rm -fr /tmp/uvcdat-build
+
+RUN useradd -d /data -m -U uvcdat
+
+VOLUME /data
+WORKDIR /data
+USER uvcdat
+
+ENV UVCDAT_SETUP_PATH /opt/uvcdat
+ENV PATH $UVCDAT_SETUP_PATH/bin:$UVCDAT_SETUP_PATH/Externals/bin:$PATH
+ENV PYTHONPATH $UVCDAT_SETUP_PATH/lib/python2.7/site-packages:$UVCDAT_SETUP_PATH/Externals/lib/python2.7/site-packages
+ENV LD_LIBRARY_PATH $UVCDAT_SETUP_PATH/lib:$UVCDAT_SETUP_PATH/Externals/lib
+ENV UVCDAT_ANONYMOUS_LOG no

docker/README.md

@@ -0,0 +1,53 @@
+UV-CDAT Docker images
+=====================
+
+Docker images of UV-CDAT are available at [docker hub](https://registry.hub.docker.com/u/uvcdat/uvcdat/).  These
+images are built with offscreen rendering and no GUI support due to limitations in docker; however, it is possible
+to render images from UV-CDAT scripts or directly using the commandline interface.  The primary motivation for
+the creation of these images is for easy deployment of visualization servers for CDATWeb, but users may find
+them useful for quick image generation or data exploration without needing to build and configure the UV-CDAT
+environment.
+
+To try using the latest version of UV-CDAT, first [install](https://docs.docker.com/compose/install/) Docker
+for your platform.  Note, Docker runs natively on Linux, but requires a lightweight virtual machine on Mac and
+Windows.  The management of this virtual machine is handled through the `boot2docker` command.  Typically,
+you will need to start up the virtual machine with `boot2docker start` and configure the environment with
+`$(boot2docker shellinit)`.  See the docker installation guide for more information.
+
+Running uvcdat once docker is set up is easy:
+```
+docker run -i -t uvcdat/uvcdat ipython
+```
+This will download the docker image, run it in a new container, and give an ipython shell inside the container.
+From here, you can import all python modules that are shipped with the CLI version of UV-CDAT.
+
+Building the docker images
+--------------------------
+
+The `Dockerfile` for building UV-CDAT is stored at the root level of this repository.  It is built on top of
+a custom Ubuntu 14.04 install containing all necessary dependencies.  Generally it will not be necessary to
+rebuild this base image, but if new packages are necessary from apt, they can be added to `docker/ubuntu/Dockerfile`
+and the image rebuilt with `docker build -t uvcdat/ubuntu`.
+
+When updating master, a new `uvcdat/uvcdat:latest` should be generated and pushed to docker hub.  The
+build should be initiated from a clean checkout of the repository to avoid adding unnecessary files to the
+image.  From the top level of the repository, issue the following command to build the image.
+```
+docker build -t uvcdat/uvcdat .
+```
+Once it is built successfully you can test the image with ctest and upload the results to
+[CDash](https://open.cdash.org/index.php?project=UV-CDAT&display=project) with
+```
+docker run ctest -S /usr/src/uvcdat/CMake/dashboard/docker.cmake -VV
+```
+Finally, if you have push access to the docker hub account, you can push the image up to docker.
+```
+docker push uvcdat/uvcdat
+```
+
+Note that UV-CDAT probably won't build with the standard VM setup by `boot2docker`.  If you get build errors, you
+may need to increase the memory and disk size when initializing.  For example, to create a VM with 4 GB of RAM and 
+50 GB of disk space:
+```
+boot2docker --memory=4096 --disksize=50000
+```

docker/ubuntu/Dockerfile

@@ -0,0 +1,8 @@
+FROM ubuntu:14.04
+MAINTAINER UV-CDAT Developers <uvcdat-support@llnl.gov>
+
+RUN apt-get update && apt-get install -y git gfortran g++ libffi-dev libsqlite-dev libssl-dev libbz2-dev libexpat-dev ncurses-dev curl make wget libjpeg-dev libpng-dev
+
+RUN curl http://www.cmake.org/files/v3.2/cmake-3.2.0-rc2-Linux-x86_64.tar.gz | tar -v -C /opt -zx
+
+ENV PATH /opt/cmake-3.2.0-rc2-Linux-x86_64/bin:$PATH