From c250aea2fe7ef1949adc8453c001c6c5c791cf64 Mon Sep 17 00:00:00 2001 From: Christoph Schranz Date: Tue, 5 Jan 2021 12:39:30 +0100 Subject: [PATCH] update docker-stacks, torchviz, dockerhub-image description --- .build/Dockerfile | 234 ++++++++++-------------------- .build/docker-stacks | 2 +- .build/fix-permissions | 28 ++-- .build/jupyter_notebook_config.py | 4 +- .build/start.sh | 30 ++-- README.md | 106 +++++++++----- generate-Dockerfile.sh | 2 +- src/Dockerfile.gpulibs | 2 + 8 files changed, 183 insertions(+), 225 deletions(-) diff --git a/.build/Dockerfile b/.build/Dockerfile index f920a0f..b01d6df 100755 --- a/.build/Dockerfile +++ b/.build/Dockerfile @@ -2,7 +2,7 @@ # **Please do not change this file directly!** # To adapt this Dockerfile, adapt 'generate-Dockerfile.sh' or 'src/Dockerfile.usefulpackages'. -# More information can be found in the documentation. +# More information can be found in the README under configuration. # Use NVIDIA CUDA as base image and run the same installation as in the other packages. @@ -22,7 +22,8 @@ RUN chmod 1777 /tmp && chmod 1777 /var/tmp # Ubuntu 20.04 (focal) # https://hub.docker.com/_/ubuntu/?tab=tags&name=focal # OS/ARCH: linux/amd64 -ARG ROOT_CONTAINER=ubuntu:focal-20200423@sha256:238e696992ba9913d24cfc3727034985abd136e08ee3067982401acdc30cbf3f +ARG ROOT_CONTAINER=ubuntu:focal-20201106@sha256:4e4bc990609ed865e07afc8427c30ffdddca5153fd4e82c20d8f0783a291e241 + LABEL maintainer="Jupyter Project " ARG NB_USER="jovyan" @@ -34,13 +35,33 @@ SHELL ["/bin/bash", "-o", "pipefail", "-c"] USER root +# ---- Miniforge installer ---- +# Default values can be overridden at build time +# (ARGS are in lower case to distinguish them from ENV) +# Check https://github.com/conda-forge/miniforge/releases +# Conda version +ARG conda_version="4.9.2" +# Miniforge installer patch version +ARG miniforge_patch_number="0" +# Miniforge installer architecture +ARG miniforge_arch="x86_64" +# Python implementation to use +# can be either Miniforge3 to use Python or Miniforge-pypy3 to use PyPy +ARG miniforge_python="Miniforge3" + +# Miniforge archive to install +ARG miniforge_version="${conda_version}-${miniforge_patch_number}" +# Miniforge installer +ARG miniforge_installer="${miniforge_python}-${miniforge_version}-Linux-${miniforge_arch}.sh" +# Miniforge checksum +ARG miniforge_checksum="6321775eb2c02d7f51d3a9004ce0be839099f126f4099c781531428536669560" + # Install all OS dependencies for notebook server that starts but lacks all # features (e.g., download as all possible file formats) ENV DEBIAN_FRONTEND noninteractive RUN apt-get update \ && apt-get install -yq --no-install-recommends \ wget \ - bzip2 \ ca-certificates \ sudo \ locales \ @@ -61,16 +82,21 @@ ENV CONDA_DIR=/opt/conda \ LANG=en_US.UTF-8 \ LANGUAGE=en_US.UTF-8 ENV PATH=$CONDA_DIR/bin:$PATH \ - HOME=/home/$NB_USER + HOME=/home/$NB_USER \ + CONDA_VERSION="${conda_version}" \ + MINIFORGE_VERSION="${miniforge_version}" # Copy a script that we will use to correct permissions after running certain commands COPY fix-permissions /usr/local/bin/fix-permissions RUN chmod a+rx /usr/local/bin/fix-permissions # Enable prompt color in the skeleton .bashrc before creating the default NB_USER -RUN sed -i 's/^#force_color_prompt=yes/force_color_prompt=yes/' /etc/skel/.bashrc +# hadolint ignore=SC2016 +RUN sed -i 's/^#force_color_prompt=yes/force_color_prompt=yes/' /etc/skel/.bashrc && \ + # Add call to conda init script see https://stackoverflow.com/a/58081608/4413446 + echo 'eval "$(command conda shell.bash hook 2> /dev/null)"' >> /etc/skel/.bashrc -# Create NB_USER wtih name jovyan user with UID=1000 and in the 'users' group +# Create NB_USER with name jovyan user with UID=1000 and in the 'users' group # and make sure these dirs are writable by the `users` group. RUN echo "auth requisite pam_deny.so" >> /etc/pam.d/su && \ sed -i.bak -e 's/^%admin/#%admin/' /etc/sudoers && \ @@ -83,42 +109,34 @@ RUN echo "auth requisite pam_deny.so" >> /etc/pam.d/su && \ fix-permissions $CONDA_DIR USER $NB_UID -WORKDIR $HOME ARG PYTHON_VERSION=default # Setup work directory for backward-compatibility -RUN mkdir /home/$NB_USER/work && \ - fix-permissions /home/$NB_USER - -# Install conda as jovyan and check the md5 sum provided on the download site -ENV MINICONDA_VERSION=4.8.2 \ - MINICONDA_MD5=87e77f097f6ebb5127c77662dfc3165e \ - CONDA_VERSION=4.8.2 +RUN mkdir "/home/$NB_USER/work" && \ + fix-permissions "/home/$NB_USER" +# Install conda as jovyan and check the sha256 sum provided on the download site WORKDIR /tmp -RUN wget --quiet https://repo.continuum.io/miniconda/Miniconda3-py37_${MINICONDA_VERSION}-Linux-x86_64.sh && \ - echo "${MINICONDA_MD5} *Miniconda3-py37_${MINICONDA_VERSION}-Linux-x86_64.sh" | md5sum -c - && \ - /bin/bash Miniconda3-py37_${MINICONDA_VERSION}-Linux-x86_64.sh -f -b -p $CONDA_DIR && \ - rm Miniconda3-py37_${MINICONDA_VERSION}-Linux-x86_64.sh && \ + +# Prerequisites installation: conda, pip, tini +RUN wget --quiet "https://github.com/conda-forge/miniforge/releases/download/${miniforge_version}/${miniforge_installer}" && \ + echo "${miniforge_checksum} *${miniforge_installer}" | sha256sum --check && \ + /bin/bash "${miniforge_installer}" -f -b -p $CONDA_DIR && \ + rm "${miniforge_installer}" && \ + # Conda configuration see https://conda.io/projects/conda/en/latest/configuration.html echo "conda ${CONDA_VERSION}" >> $CONDA_DIR/conda-meta/pinned && \ - conda config --system --prepend channels conda-forge && \ conda config --system --set auto_update_conda false && \ conda config --system --set show_channel_urls true && \ - conda config --system --set channel_priority strict && \ if [ ! $PYTHON_VERSION = 'default' ]; then conda install --yes python=$PYTHON_VERSION; fi && \ conda list python | grep '^python ' | tr -s ' ' | cut -d '.' -f 1,2 | sed 's/$/.*/' >> $CONDA_DIR/conda-meta/pinned && \ - conda install --quiet --yes conda && \ - conda install --quiet --yes pip && \ + conda install --quiet --yes \ + "conda=${CONDA_VERSION}" \ + 'pip' \ + 'tini=0.18.0' && \ conda update --all --quiet --yes && \ - conda clean --all -f -y && \ - rm -rf /home/$NB_USER/.cache/yarn && \ - fix-permissions $CONDA_DIR && \ - fix-permissions /home/$NB_USER - -# Install Tini -RUN conda install --quiet --yes 'tini=0.18.0' && \ conda list tini | grep tini | tr -s ' ' | cut -d ' ' -f 1,2 >> $CONDA_DIR/conda-meta/pinned && \ conda clean --all -f -y && \ + rm -rf /home/$NB_USER/.cache/yarn && \ fix-permissions $CONDA_DIR && \ fix-permissions /home/$NB_USER @@ -129,9 +147,9 @@ RUN conda install --quiet --yes 'tini=0.18.0' && \ # Do all this in a single RUN command to avoid duplicating all of the # files across image layers when the permissions change RUN conda install --quiet --yes \ - 'notebook=6.0.3' \ - 'jupyterhub=1.1.0' \ - 'jupyterlab=2.1.3' && \ + 'notebook=6.1.6' \ + 'jupyterhub=1.3.0' \ + 'jupyterlab=2.2.9' && \ conda clean --all -f -y && \ npm cache clean --force && \ jupyter notebook --generate-config && \ @@ -148,10 +166,16 @@ CMD ["start-notebook.sh"] # Copy local files as late as possible to avoid cache busting COPY start.sh start-notebook.sh start-singleuser.sh /usr/local/bin/ +# Currently need to have both jupyter_notebook_config and jupyter_server_config to support classic and lab COPY jupyter_notebook_config.py /etc/jupyter/ # Fix permissions on /etc/jupyter as root USER root + +# Prepare upgrade to JupyterLab V3.0 #1205 +RUN sed -re "s/c.NotebookApp/c.ServerApp/g" \ + /etc/jupyter/jupyter_notebook_config.py > /etc/jupyter/jupyter_server_config.py + RUN fix-permissions /etc/jupyter/ # Switch back to jovyan to avoid accidental container runs as root @@ -173,29 +197,27 @@ USER root # Install all OS dependencies for fully functional notebook server RUN apt-get update && apt-get install -yq --no-install-recommends \ build-essential \ - emacs-nox \ vim-tiny \ git \ inkscape \ - jed \ libsm6 \ libxext-dev \ libxrender1 \ lmodern \ netcat \ - python-dev \ # ---- nbconvert dependencies ---- texlive-xetex \ texlive-fonts-recommended \ texlive-plain-generic \ - # Optional dependency - texlive-fonts-extra \ # ---- tzdata \ unzip \ - nano \ + nano-tiny \ && apt-get clean && rm -rf /var/lib/apt/lists/* +# Create alternative for nano -> nano-tiny +RUN update-alternatives --install /usr/bin/nano nano /bin/nano-tiny 10 + # Switch back to jovyan to avoid accidental container runs as root USER $NB_UID @@ -210,10 +232,10 @@ LABEL maintainer="Jupyter Project " USER root -# ffmpeg for matplotlib anim & dvipng for latex labels +# ffmpeg for matplotlib anim & dvipng+cm-super for latex labels RUN apt-get update && \ - apt-get install -y --no-install-recommends ffmpeg dvipng && \ - rm -rf /var/lib/apt/lists/* + apt-get install -y --no-install-recommends ffmpeg dvipng cm-super && \ + apt-get clean && rm -rf /var/lib/apt/lists/* USER $NB_UID @@ -221,31 +243,29 @@ USER $NB_UID RUN conda install --quiet --yes \ 'beautifulsoup4=4.9.*' \ 'conda-forge::blas=*=openblas' \ - 'bokeh=2.0.*' \ + 'bokeh=2.2.*' \ 'bottleneck=1.3.*' \ - 'cloudpickle=1.4.*' \ + 'cloudpickle=1.6.*' \ 'cython=0.29.*' \ - 'dask=2.15.*' \ + 'dask=2020.12.*' \ 'dill=0.3.*' \ - 'h5py=2.10.*' \ - 'hdf5=1.10.*' \ - 'ipywidgets=7.5.*' \ + 'h5py=3.1.*' \ + 'ipywidgets=7.6.*' \ 'ipympl=0.5.*'\ - 'matplotlib-base=3.2.*' \ - # numba update to 0.49 fails resolving deps. - 'numba=0.48.*' \ + 'matplotlib-base=3.3.*' \ + 'numba=0.52.*' \ 'numexpr=2.7.*' \ - 'pandas=1.0.*' \ + 'pandas=1.1.*' \ 'patsy=0.5.*' \ - 'protobuf=3.11.*' \ + 'protobuf=3.14.*' \ 'pytables=3.6.*' \ - 'scikit-image=0.16.*' \ - 'scikit-learn=0.22.*' \ - 'scipy=1.4.*' \ - 'seaborn=0.10.*' \ + 'scikit-image=0.18.*' \ + 'scikit-learn=0.24.*' \ + 'scipy=1.5.*' \ + 'seaborn=0.11.*' \ 'sqlalchemy=1.3.*' \ - 'statsmodels=0.11.*' \ - 'sympy=1.5.*' \ + 'statsmodels=0.12.*' \ + 'sympy=1.7.*' \ 'vincent=0.4.*' \ 'widgetsnbextension=3.5.*'\ 'xlrd=1.2.*' \ @@ -283,102 +303,6 @@ RUN MPLBACKEND=Agg python -c "import matplotlib.pyplot" && \ USER $NB_UID -WORKDIR $HOME - - ############################################################################ - ################ Dependency: jupyter/datascience-notebook ################## - ############################################################################ - -# Copyright (c) Jupyter Development Team. -# Distributed under the terms of the Modified BSD License. - -LABEL maintainer="Jupyter Project " - -# Set when building on Travis so that certain long-running build steps can -# be skipped to shorten build time. -ARG TEST_ONLY_BUILD - -# Fix DL4006 -SHELL ["/bin/bash", "-o", "pipefail", "-c"] - -USER root - -# R pre-requisites -RUN apt-get update && \ - apt-get install -y --no-install-recommends \ - fonts-dejavu \ - gfortran \ - gcc && \ - rm -rf /var/lib/apt/lists/* - -# Julia dependencies -# install Julia packages in /opt/julia instead of $HOME -ENV JULIA_DEPOT_PATH=/opt/julia -ENV JULIA_PKGDIR=/opt/julia -ENV JULIA_VERSION=1.4.1 - -WORKDIR /tmp - -# hadolint ignore=SC2046 -RUN mkdir "/opt/julia-${JULIA_VERSION}" && \ - wget -q https://julialang-s3.julialang.org/bin/linux/x64/$(echo "${JULIA_VERSION}" | cut -d. -f 1,2)"/julia-${JULIA_VERSION}-linux-x86_64.tar.gz" && \ - echo "fd6d8cadaed678174c3caefb92207a3b0e8da9f926af6703fb4d1e4e4f50610a *julia-${JULIA_VERSION}-linux-x86_64.tar.gz" | sha256sum -c - && \ - tar xzf "julia-${JULIA_VERSION}-linux-x86_64.tar.gz" -C "/opt/julia-${JULIA_VERSION}" --strip-components=1 && \ - rm "/tmp/julia-${JULIA_VERSION}-linux-x86_64.tar.gz" -RUN ln -fs /opt/julia-*/bin/julia /usr/local/bin/julia - -# Show Julia where conda libraries are \ -RUN mkdir /etc/julia && \ - echo "push!(Libdl.DL_LOAD_PATH, \"$CONDA_DIR/lib\")" >> /etc/julia/juliarc.jl && \ - # Create JULIA_PKGDIR \ - mkdir "${JULIA_PKGDIR}" && \ - chown "${NB_USER}" "${JULIA_PKGDIR}" && \ - fix-permissions "${JULIA_PKGDIR}" - -USER $NB_UID - -# R packages including IRKernel which gets installed globally. -RUN conda install --quiet --yes \ - 'r-base=3.6.3' \ - 'r-caret=6.0*' \ - 'r-crayon=1.3*' \ - 'r-devtools=2.3*' \ - 'r-forecast=8.12*' \ - 'r-hexbin=1.28*' \ - 'r-htmltools=0.4*' \ - 'r-htmlwidgets=1.5*' \ - 'r-irkernel=1.1*' \ - 'r-nycflights13=1.0*' \ - 'r-plyr=1.8*' \ - 'r-randomforest=4.6*' \ - 'r-rcurl=1.98*' \ - 'r-reshape2=1.4*' \ - 'r-rmarkdown=2.1*' \ - 'r-rsqlite=2.2*' \ - 'r-shiny=1.4*' \ - 'r-tidyverse=1.3*' \ - 'rpy2=3.1*' \ - && \ - conda clean --all -f -y && \ - fix-permissions "${CONDA_DIR}" && \ - fix-permissions "/home/${NB_USER}" - -# Add Julia packages. Only add HDF5 if this is not a test-only build since -# it takes roughly half the entire build time of all of the images on Travis -# to add this one package and often causes Travis to timeout. -# -# Install IJulia as jovyan and then move the kernelspec out -# to the system share location. Avoids problems with runtime UID change not -# taking effect properly on the .local folder in the jovyan home dir. -RUN julia -e 'import Pkg; Pkg.update()' && \ - (test $TEST_ONLY_BUILD || julia -e 'import Pkg; Pkg.add("HDF5")') && \ - julia -e "using Pkg; pkg\"add IJulia\"; pkg\"precompile\"" && \ - # move kernelspec out of home \ - mv "${HOME}/.local/share/jupyter/kernels/julia"* "${CONDA_DIR}/share/jupyter/kernels/" && \ - chmod -R go+rx "${CONDA_DIR}/share/jupyter" && \ - rm -rf "${HOME}/.local" && \ - fix-permissions "${JULIA_PKGDIR}" "${CONDA_DIR}/share/jupyter" - WORKDIR $HOME ############################################################################ @@ -403,6 +327,8 @@ RUN conda install --quiet --yes \ torchvision \ cudatoolkit=10.1 -c pytorch # pip install torch_nightly -f https://download.pytorch.org/whl/nightly/cu90/torch_nightly.html && \ +RUN pip install --no-cache-dir torchviz + # Clean installation RUN conda clean --all -f -y && \ diff --git a/.build/docker-stacks b/.build/docker-stacks index d676cdf..703d8b2 160000 --- a/.build/docker-stacks +++ b/.build/docker-stacks @@ -1 +1 @@ -Subproject commit d676cdf9b4847b5a5da2d4367aed56265174a5ef +Subproject commit 703d8b2dcb886be2fe5aa4660a48fbcef647e7aa diff --git a/.build/fix-permissions b/.build/fix-permissions index 659b276..3b038f2 100755 --- a/.build/fix-permissions +++ b/.build/fix-permissions @@ -18,18 +18,18 @@ set -e for d in "$@"; do - find "$d" \ - ! \( \ - -group $NB_GID \ - -a -perm -g+rwX \ - \) \ - -exec chgrp $NB_GID {} \; \ - -exec chmod g+rwX {} \; - # setuid,setgid *on directories only* - find "$d" \ - \( \ - -type d \ - -a ! -perm -6000 \ - \) \ - -exec chmod +6000 {} \; + find "$d" \ + ! \( \ + -group $NB_GID \ + -a -perm -g+rwX \ + \) \ + -exec chgrp $NB_GID {} \; \ + -exec chmod g+rwX {} \; + # setuid, setgid *on directories only* + find "$d" \ + \( \ + -type d \ + -a ! -perm -6000 \ + \) \ + -exec chmod +6000 {} \; done diff --git a/.build/jupyter_notebook_config.py b/.build/jupyter_notebook_config.py index 15ac96a..19b5e8b 100755 --- a/.build/jupyter_notebook_config.py +++ b/.build/jupyter_notebook_config.py @@ -7,7 +7,7 @@ import os import errno import stat -c = get_config() +c = get_config() # noqa: F821 c.NotebookApp.ip = '0.0.0.0' c.NotebookApp.port = 8888 c.NotebookApp.open_browser = False @@ -52,4 +52,4 @@ distinguished_name = req_distinguished_name # Change default umask for all subprocesses of the notebook server if set in # the environment if 'NB_UMASK' in os.environ: - os.umask(int(os.environ['NB_UMASK'], 8)) \ No newline at end of file + os.umask(int(os.environ['NB_UMASK'], 8)) diff --git a/.build/start.sh b/.build/start.sh index b293197..17b0ec7 100755 --- a/.build/start.sh +++ b/.build/start.sh @@ -47,19 +47,6 @@ if [ $(id -u) == 0 ] ; then usermod -d /home/$NB_USER -l $NB_USER jovyan fi - # Handle case where provisioned storage does not have the correct permissions by default - # Ex: default NFS/EFS (no auto-uid/gid) - if [[ "$CHOWN_HOME" == "1" || "$CHOWN_HOME" == 'yes' ]]; then - echo "Changing ownership of /home/$NB_USER to $NB_UID:$NB_GID with options '${CHOWN_HOME_OPTS}'" - chown $CHOWN_HOME_OPTS $NB_UID:$NB_GID /home/$NB_USER - fi - if [ ! -z "$CHOWN_EXTRA" ]; then - for extra_dir in $(echo $CHOWN_EXTRA | tr ',' ' '); do - echo "Changing ownership of ${extra_dir} to $NB_UID:$NB_GID with options '${CHOWN_EXTRA_OPTS}'" - chown $CHOWN_EXTRA_OPTS $NB_UID:$NB_GID $extra_dir - done - fi - # handle home and working directory if the username changed if [[ "$NB_USER" != "jovyan" ]]; then # changing username, make sure homedir exists @@ -76,11 +63,24 @@ if [ $(id -u) == 0 ] ; then fi fi + # Handle case where provisioned storage does not have the correct permissions by default + # Ex: default NFS/EFS (no auto-uid/gid) + if [[ "$CHOWN_HOME" == "1" || "$CHOWN_HOME" == 'yes' ]]; then + echo "Changing ownership of /home/$NB_USER to $NB_UID:$NB_GID with options '${CHOWN_HOME_OPTS}'" + chown $CHOWN_HOME_OPTS $NB_UID:$NB_GID /home/$NB_USER + fi + if [ ! -z "$CHOWN_EXTRA" ]; then + for extra_dir in $(echo $CHOWN_EXTRA | tr ',' ' '); do + echo "Changing ownership of ${extra_dir} to $NB_UID:$NB_GID with options '${CHOWN_EXTRA_OPTS}'" + chown $CHOWN_EXTRA_OPTS $NB_UID:$NB_GID $extra_dir + done + fi + # Change UID:GID of NB_USER to NB_UID:NB_GID if it does not match if [ "$NB_UID" != $(id -u $NB_USER) ] || [ "$NB_GID" != $(id -g $NB_USER) ]; then echo "Set user $NB_USER UID:GID to: $NB_UID:$NB_GID" if [ "$NB_GID" != $(id -g $NB_USER) ]; then - groupadd -g $NB_GID -o ${NB_GROUP:-${NB_USER}} + groupadd -f -g $NB_GID -o ${NB_GROUP:-${NB_USER}} fi userdel $NB_USER useradd --home /home/$NB_USER -u $NB_UID -g $NB_GID -G 100 -l $NB_USER @@ -101,7 +101,7 @@ if [ $(id -u) == 0 ] ; then echo "Executing the command: ${cmd[@]}" exec sudo -E -H -u $NB_USER PATH=$PATH XDG_CACHE_HOME=/home/$NB_USER/.cache PYTHONPATH=${PYTHONPATH:-} "${cmd[@]}" else - if [[ "$NB_UID" == "$(id -u jovyan)" && "$NB_GID" == "$(id -g jovyan)" ]]; then + if [[ "$NB_UID" == "$(id -u jovyan 2>/dev/null)" && "$NB_GID" == "$(id -g jovyan 2>/dev/null)" ]]; then # User is not attempting to override user/group via environment # variables, but they could still have overridden the uid/gid that # container runs as. Check that the user has an entry in the passwd diff --git a/README.md b/README.md index 6c6b2ef..4b2939f 100644 --- a/README.md +++ b/README.md @@ -21,62 +21,81 @@ The image of this repository is available on [Dockerhub](https://hub.docker.com/ ## Requirements -1. A NVIDIA GPU +1. A computer with a NVIDIA GPU 2. Install [Docker](https://www.docker.com/community-edition#/download) version **1.10.0+** and [Docker Compose](https://docs.docker.com/compose/install/) version **1.6.0+**. -3. Get access to your GPU via CUDA drivers within Docker containers. Therfore, check out this -[medium article](https://medium.com/@christoph.schranz/set-up-your-own-gpu-based-jupyterlab-e0d45fcacf43). - The CUDA toolkit is not required on the host system, as it will be deployed - in [NVIDIA-docker](https://github.com/NVIDIA/nvidia-docker). +3. Get access to your GPU via CUDA drivers within Docker containers. You can be sure that you can access your GPU within Docker, - if the command `docker run --gpus all nvidia/cuda:10.1-base-ubuntu18.04 nvidia-smi` + if the command `docker run --gpus all nvidia/cuda:10.1-cudnn7-runtime-ubuntu18.04 nvidia-smi` returns a result similar to this one: ```bash - Mon Jun 22 09:06:28 2020 + Tue Jan 5 09:38:21 2021 +-----------------------------------------------------------------------------+ - | NVIDIA-SMI 440.82 Driver Version: 440.82 CUDA Version: 10.1 | + | NVIDIA-SMI 450.80.02 Driver Version: 450.80.02 CUDA Version: 10.1 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | + | | | MIG M. | |===============================+======================+======================| | 0 GeForce RTX 207... Off | 00000000:01:00.0 On | N/A | - | 0% 46C P8 9W / 215W | 424MiB / 7974MiB | 6% Default | + | 0% 40C P8 7W / 215W | 360MiB / 7974MiB | 1% Default | + | | | N/A | +-------------------------------+----------------------+----------------------+ +-----------------------------------------------------------------------------+ - | Processes: GPU Memory | - | GPU PID Type Process name Usage | + | Processes: | + | GPU GI CI PID Type Process name GPU Memory | + | ID ID Usage | |=============================================================================| +-----------------------------------------------------------------------------+ - ``` -4. Clone the Repository or pull the image from - [Dockerhub](https://hub.docker.com/repository/docker/cschranz/gpu-jupyter): - ```bash - git clone https://github.com/iot-salzburg/gpu-jupyter.git - cd gpu-jupyter - ``` + ``` + If you don't get an output similar than this one, follow the installation steps in this +[medium article](https://medium.com/@christoph.schranz/set-up-your-own-gpu-based-jupyterlab-e0d45fcacf43). + The CUDA toolkit is not required on the host system, as it will be + installed within the Docker containers using [NVIDIA-docker](https://github.com/NVIDIA/nvidia-docker). + It is also important to keep your installed CUDA version in mind, when you pull images. + **You can't run images based on `nvidia/cuda:11.1` if you have only CUDA version 10.1 installed.** + Check your host's CUDA-version with `nvcc --version` and update to at least + the same version you want to pull. + +4. Pull and run the image. This can last some hours, as a whole data-science + environment will be downloaded: + ```bash + cd your-working-directory + docker run --gpus all -d -it -p 8848:8888 -v data:/home/jovyan/work -e GRANT_SUDO=yes -e JUPYTER_ENABLE_LAB=yes --user root cschranz/gpu-jupyter:v1.1_cuda-10.1_ubuntu-18.04_python-only + ``` + This starts a new instance of the GPU-Jupyter service on at [http://localhost:8848](http://localhost:8848) (port `8484`). + The default password is `asdf` which should be changed as described [below](#set-password). + Furthermore, data within the host's `data` directory is shared with the container. + +Within the Jupyterlab instance, you can check if you can access your GPU by opening a new terminal window and running +`nvidia-smi`. In terminal windows, you can also install new packages for your own projects. +Some example code can be found in the repository under `extra/Getting_Started`. +If you want to learn more about Jupyterlab, check out this [tutorial](https://www.youtube.com/watch?v=7wfPqAyYADY). -## Quickstart -First of all, it is necessary to generate the `Dockerfile` that is based on +## Build a modified version + +First, it is necessary to generate the `Dockerfile` in `.build`, that is based on the NIVIDA base image and the [docker-stacks](https://github.com/jupyter/docker-stacks). As soon as you have access to your GPU within Docker containers -(make sure the command `docker run --gpus all nvidia/cuda:10.1-base-ubuntu18.04 nvidia-smi` +(make sure the command `docker run --gpus all nvidia/cuda:10.1-cudnn7-runtime-ubuntu18.04 nvidia-smi` shows your GPU statistics), you can generate the Dockerfile, build and run it. The following commands will start *GPU-Jupyter* on [localhost:8848](http://localhost:8848) with the default password `asdf`. ```bash + git clone https://github.com/iot-salzburg/gpu-jupyter.git + cd gpu-jupyter # generate a Dockerfile with python and without Julia and R ./generate-Dockerfile.sh --no-datascience-notebook docker build -t gpu-jupyter .build/ # will take a while - docker run -d -p [port]:8888 gpu-jupyter # starts gpu-jupyter WITHOUT GPU support + docker run --gpus all -d -it -p 8848:8888 -v $pwd/data:/home/jovyan/work -e GRANT_SUDO=yes -e JUPYTER_ENABLE_LAB=yes --user root --restart always --name gpu-jupyter_1 gpu-jupyter ``` -To run the container WITH GPU support, a local data volume and some other configurations, run: - ```bash -docker run --gpus all -d -it -p 8848:8888 -v $(pwd)/data:/home/jovyan/work -e GRANT_SUDO=yes -e JUPYTER_ENABLE_LAB=yes --user root --restart always --name gpu-jupyter_1 gpu-jupyter -``` +This starts a container WITH GPU support, a shared local data volume `data` +and some other configurations like root permissions which are necessary to install packages within the container. +For more configurations, scroll down to [Configuration of the Dockerfile-Generation](#configuration-of-the-dockerfile-generation). ### Start via Docker Compose @@ -92,14 +111,16 @@ underlying `docker-compose.yml`: With these commands we can see if everything worked well: ```bash -bash show-local.sh # a env-var safe wrapper for 'docker-compose logs -f' docker ps -docker logs [service-name] +docker logs [service-name] # or +bash show-local.sh # a env-var safe wrapper for 'docker-compose logs -f' ``` In order to stop the local deployment, run: ```bash +docker ps +docker rm -f [service-name] # or ./stop-local.sh ``` @@ -134,13 +155,16 @@ the essential `gpulibs` are installed, but not the packages within `src/Dockerfi ### Custom Installations -**As `.build/Dockerfile` is overwritten, it is suggested to append custom installations either +Custom packages can be installed within a container, or by modifying the file +`src/Dockerfile.usefulpackages`. +**As `.build/Dockerfile` is overwritten each time a Dockerfile is generated, +it is suggested to append custom installations either within `src/Dockerfile.usefulpackages` or in `generate-Dockerfile.sh`.** -If you think some package is missing in the default stack, please let us know! +If an essential package is missing in the default stack, please let us know! -### Set the Password +### Set Password Please set a new password using `src/jupyter_notebook_config.json`. Therefore, hash your password in the form (password)(salt) using a sha1 hash generator, e.g., the sha1 generator of [sha1-online.com](http://www.sha1-online.com/). @@ -162,24 +186,30 @@ Then update the config file as shown below and restart the service. #### Update CUDA to another version -Please check version compatibilities for [CUDA and Pytorch](https://pytorch.org/get-started/locally/) - respectively [CUDA and Tensorflow](https://www.tensorflow.org/install/gpu) previously. -To update CUDA to another version, change in `Dockerfile.header` +The host's CUDA-version must be equal or higher than that of the +container itself (in `Dockerfile.header`). +Check the host's version with `nvcc --version` and the version compatibilities +for CUDA-dependent packages as [Pytorch](https://pytorch.org/get-started/locally/) + respectively [Tensorflow](https://www.tensorflow.org/install/gpu) previously. +Then modify, if supported, the CUDA-version in `Dockerfile.header` to, e.g.: the line: - FROM nvidia/cuda:10.1-base-ubuntu18.04 + FROM nvidia/cuda:11.1-base-ubuntu20.04 and in the `Dockerfile.pytorch` the line: - cudatoolkit=10.1 + cudatoolkit=11.1 -Then re-generate and re-run the image, as closer described above: +Then re-generate, re-build and run the updated image, as closer described above: +Note that a change in the first line of the Dockerfile will re-build the whole image. ```bash ./generate-Dockerfile.sh -./start-local.sh -p 8848 +docker build -t gpu-jupyter .build/ # will take a while +docker run --gpus all -d -it -p 8848:8888 -v $pwd/data:/home/jovyan/work -e GRANT_SUDO=yes -e JUPYTER_ENABLE_LAB=yes --user root --restart always --name gpu-jupyter_1 gpu-jupyter ``` + #### Update Docker-Stack The [docker-stacks](https://github.com/jupyter/docker-stacks) are used as a diff --git a/generate-Dockerfile.sh b/generate-Dockerfile.sh index d9916fc..7dd5255 100755 --- a/generate-Dockerfile.sh +++ b/generate-Dockerfile.sh @@ -5,7 +5,7 @@ cd $(cd -P -- "$(dirname -- "$0")" && pwd -P) export DOCKERFILE=".build/Dockerfile" export STACKS_DIR=".build/docker-stacks" # please test the build of the commit in https://github.com/jupyter/docker-stacks/commits/master in advance -export HEAD_COMMIT="04f7f60d34a674a2964d96a6cb97c57a7870a828" +export HEAD_COMMIT="703d8b2dcb886be2fe5aa4660a48fbcef647e7aa" while [[ "$#" -gt 0 ]]; do case $1 in -c|--commit) HEAD_COMMIT="$2"; shift;; diff --git a/src/Dockerfile.gpulibs b/src/Dockerfile.gpulibs index ebd5cce..b933e05 100644 --- a/src/Dockerfile.gpulibs +++ b/src/Dockerfile.gpulibs @@ -16,6 +16,8 @@ RUN conda install --quiet --yes \ torchvision \ cudatoolkit=10.1 -c pytorch # pip install torch_nightly -f https://download.pytorch.org/whl/nightly/cu90/torch_nightly.html && \ +RUN pip install --no-cache-dir torchviz + # Clean installation RUN conda clean --all -f -y && \