update

This reverts commit c7bff5448a, reversing
changes made to cea7bc59f3.

.gitignore

@@ -1,3 +1,2 @@
 /docker-stacks
-jupyterhub-data
 *.tar

Dockerfile

@@ -1,4 +1,3 @@
-# syntax=docker/dockerfile:1
 FROM debian:bullseye
 ARG CREATE_USER="jovyan"
 ARG CREATE_USER_PASSWD="password"

Dockerfile.jupyterhub

@@ -8,8 +8,6 @@ ARG ROOT_PASSWD="password"
 ARG HOME="/home/${CREATE_USER}"
 ARG DEBIAN_FRONTEND="noninteractive"
 ENV DEBIAN_FRONTEND=${DEBIAN_FRONTEND}
-ARG NODEJS_VERSION='18'
-ENV NODEJS_VERSION=${NODEJS_VERSION}
 
 USER root
 RUN <<EOT
@@ -21,8 +19,6 @@ apt-get install -y tzdata
 ln -sf /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
 echo 'Asia/Shanghai' > /etc/timezone
 dpkg-reconfigure -f noninteractive tzdata
-sudo apt-get remove --purge libnode72:amd64 -y
-curl -fsSL https://deb.nodesource.com/setup_${NODEJS_VERSION}.x | sudo -E bash - 
 # 安装所需的软件包
 apt-get install -y python3 python3-pip gcc g++ build-essential nodejs npm gdebi-core curl wget openssh-server vim lrzsz net-tools sudo git
 # 创建新用户
@@ -33,26 +29,17 @@ EOT
 
 RUN <<EOT
 #!/bin/bash
+mkdir -p ~/.pip
+echo "
+[global]
+index-url = https://mirrors.aliyun.com/pypi/simple/
+
+[install]
+trusted-host=mirrors.aliyun.com
+" >> ~/.pip/pip.conf
 # 安装 Jupyter 和相关软件
-npm install -g configurable-http-proxy yarn --registry=https://registry.npmmirror.com
-python3 -m pip install ipython jupyter_packaging jupyterhub jupyterlab notebook radian pycurl jupyter-rsession-proxy \
-ipykernel jupyterlab-language-pack-zh-CN jupyterlab-git jupyterlab-system-monitor jupyter_nbextensions_configurator \
-jupyter_contrib_nbextensions jupyterlab-unfold jupyterlab_widgets jupyterlab-drawio jupyterlab-spreadsheet-editor \
-jupyterlab-cell-flash jedi-language-server jupyterlab_code_formatter jupyterlab-spellchecker jupyterlab_vim nbresuse \
-ipydrawio jedi ipympl black isort theme-darcula ipywidgets tensorboard jupyterlab_latex jupyter_bokeh autopep8 \
-xeus-python jupyterlab-lsp python-lsp-server dockerspawner jupyterhub-nativeauthenticator lckr_jupyterlab_variableinspector -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
-git clone https://ghproxy.dockless.eu.org/https://github.com/arose/nglview
-cd nglview
-python3 setup.py install
-cd js
-rm -rf node_modules package-lock.json
-npm install typescript@latest --registry=https://registry.npmmirror.com
-npm install --registry=https://registry.npmmirror.com
-cd ../..
-python -m ipykernel install --sys-prefix
-jupyter nbextension enable --py --sys-prefix widgetsnbextension
-jupyter nbextension enable --py --sys-prefix nglview
-jupyter-nbextension enable nglview --py --sys-prefix
+npm install -g configurable-http-proxy
+python3 -m pip install jupyterhub jupyterlab notebook radian pycurl jupyter-rsession-proxy ipykernel jupyterlab-language-pack-zh-CN jupyterlab-git jupyterlab-system-monitor jupyter_nbextensions_configurator jupyter_contrib_nbextensions jupyterlab-unfold jupyterlab_widgets jupyterlab-drawio jupyterlab-spreadsheet-editor jupyterlab-cell-flash jedi-language-server jupyterlab_code_formatter jupyterlab-spellchecker jupyterlab_vim nbresuse ipydrawio jedi ipympl black isort theme-darcula ipywidgets tensorboard jupyterlab_latex jupyter_bokeh autopep8 xeus-python jupyterlab-lsp python-lsp-server nglview dockerspawner jupyterhub-nativeauthenticator lckr_jupyterlab_variableinspector
 # 创建 JupyterHub 配置目录
 mkdir -p /root/.jupyterhub
 EOT
@@ -73,9 +60,22 @@ sudo wget "https://download2.rstudio.org/server/$(lsb_release -cs)/amd64/rstudio
 sudo chmod +x /tmp/rstudio-server.deb 
 sudo gdebi -n /tmp/rstudio-server.deb
 sudo rm -rf /tmp/rstudio-server.deb
+EOT
+
+RUN <<EOT
+#!/bin/bash
+# 安装micromamba
+echo "1" | bash <(curl -s https://raw.githubusercontent.com/hotwa/MicroMamba_Installer/main/install.sh)
+mkdir -p /root/workspace/personal /root/workspace/project
+EOT
+
+RUN <<EOT
+#!/bin/bash
 # 清理和减小镜像大小
 apt-get clean 
 rm -rf /var/lib/apt/lists/* /tmp/* /var/tmp/* /usr/share/man/?? /usr/share/man/??_*
 EOT
 
+
+
 CMD ["jupyterhub", "-f", "/srv/jupyterhub/jupyterhub_config.py"]

docker-compose.yml

@@ -12,7 +12,7 @@ services:
       args:
         JUPYTERHUB_VERSION: latest
     restart: always
-    image: quay.io/hotwa/jupyterhub:latest
+    image: hotwa/jupyterhub # registry.cn-hangzhou.aliyuncs.com/hotwa/jupyterhub
     container_name: jupyterhub
     networks:
       - jupyterhub-network
@@ -23,17 +23,16 @@ services:
       # within the container
       - "/var/run/docker.sock:/var/run/docker.sock:rw"
       # Bind Docker volume on host for JupyterHub database and cookie secrets
-      - "./jupyterhub-data:/data"
+      - "jupyterhub-data:/data"
     ports:
       - "8000:8000"
-      - "8080:8080"
     environment:
       # This username will be a JupyterHub admin
       JUPYTERHUB_ADMIN: admin
       # All containers will join this network
       DOCKER_NETWORK_NAME: jupyterhub-network
       # JupyterHub will spawn this Notebook image for users
-      DOCKER_NOTEBOOK_IMAGE: quay.io/hotwa/notebook:latest
+      DOCKER_NOTEBOOK_IMAGE: hotwa/notebook:latest
       # Notebook directory inside user image
       DOCKER_NOTEBOOK_DIR: /home/jovyan/work
 

id_rsa

@@ -0,0 +1,49 @@
+-----BEGIN OPENSSH PRIVATE KEY-----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+-----END OPENSSH PRIVATE KEY-----

jupyterhub_config.py

@@ -10,21 +10,7 @@ c = get_config()  # noqa: F821
 # avoid having to rebuild the JupyterHub container every time we change a
 # configuration parameter.
 
-# from dockerspawner import DockerSpawner
-
-# class MyDockerSpawner(DockerSpawner):
-#     def start(self):
-#         # 启动父类的start方法
-#         self.user_options['environment']['JUPYTER_ENABLE_NBEXTENSIONS'] = 'true'
-#         self.user_options['cmd'] = [
-#             'bash',
-#             '-c',
-#             'pip install nglview jupyter_packaging -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com && jupyter nbextension enable nglview --py --sys-prefix && jupyter labextension install nglview-js-widgets && jupyter labextension install @jupyter-widgets/jupyterlab-manager && start-singleuser.sh'
-#         ]
-#         return super().start()
-        
 # Spawn single-user servers as Docker containers
-c.Authenticator.allow_all = True
 c.JupyterHub.spawner_class = "dockerspawner.DockerSpawner"
 
 # Spawn containers from this image

nginx.conf

@@ -6,7 +6,7 @@ http {
 
     map $http_upgrade $connection_upgrade {
         default upgrade;
-        '' close;
+        ''      close;
     }
 
     server {
@@ -33,7 +33,6 @@ http {
             proxy_set_header Host $host;
             proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;
             proxy_set_header X-Forwarded-Proto $scheme;
-            proxy_set_header X-Forwarded-Port $server_port; # 添加这一行,保持原有的端口号
 
             # WebSocket support
             proxy_http_version 1.1;

spawnerdockerfile/Dockerfile.base-notebook

@@ -78,8 +78,6 @@ ARG ROOT_PASSWD="password"
 ARG HOME="/home/${CREATE_USER}"
 ARG DEBIAN_FRONTEND="noninteractive"
 ENV DEBIAN_FRONTEND=${DEBIAN_FRONTEND}
-ARG NODEJS_VERSION='18'
-ENV NODEJS_VERSION=${NODEJS_VERSION}
 
 COPY install.sh /tmp
 RUN <<EOT
@@ -92,10 +90,7 @@ ln -sf /usr/share/zoneinfo/Asia/Shanghai /etc/localtime
 echo 'Asia/Shanghai' > /etc/timezone
 dpkg-reconfigure -f noninteractive tzdata
 # 安装所需的软件包
-sudo apt-get remove --purge libnode72:amd64 -y
-curl -fsSL https://deb.nodesource.com/setup_${NODEJS_VERSION}.x | sudo -E bash - 
 apt-get install -y python3 python3-pip gcc g++ build-essential nodejs npm gdebi-core curl wget openssh-server vim lrzsz net-tools sudo git nodejs
-npm install -g configurable-http-proxy yarn --registry=https://registry.npmmirror.com
 # 创建新用户
 useradd -m -s /bin/bash ${CREATE_USER}
 echo "${CREATE_USER}:${CREATE_USER_PASSWD}" | chpasswd
@@ -105,7 +100,11 @@ EOT
 # 安装 Jupyter 和相关软件
 RUN <<EOT
 #!/bin/bash
-python3 -m pip install ipython jupyterhub jupyterlab notebook jupyter_packaging -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
+npm install -g configurable-http-proxy
+git clone https://github.com/arose/nglview
+cd nglview
+python setup.py install
+cd ..
 python3 -m pip install aiohttp -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
 python3 -m pip install jupyterhub jupyterlab notebook radian pycurl aiohttp jupyter-rsession-proxy ipykernel>=6.25.0 jupyterlab-language-pack-zh-CN jupyterlab-git jupyterlab-system-monitor jupyter_nbextensions_configurator jupyter_contrib_nbextensions jupyterlab_widgets jupyterlab-drawio jupyterlab-spreadsheet-editor jupyterlab-cell-flash jedi-language-server jupyterlab_code_formatter jupyterlab-spellchecker jupyterlab_vim nbresuse ipydrawio jedi ipympl black isort theme-darcula ipywidgets tensorboard jupyterlab_latex jupyter_bokeh autopep8 xeus-python jupyterlab-lsp python-lsp-server dockerspawner jupyterhub-nativeauthenticator lckr_jupyterlab_variableinspector -i http://mirrors.aliyun.com/pypi/simple/ --trusted-host mirrors.aliyun.com
 EOT
@@ -139,58 +138,23 @@ USER ${NB_UID}
 ARG MODULAR_HOME="/home/${CREATE_USER}/.modular"
 ENV MODULAR_HOME=$MODULAR_HOME
 ENV PATH="$MODULAR_HOME/pkg/packages.modular.com_mojo/bin:/home/${CREATE_USER}/.local/bin:$PATH"
-ARG HTTP_PROXY=""
-ARG HTTPS_PROXY=""
-ARG NO_PROXY="localhost,127.0.0.1"
-ENV HTTP_PROXY=$HTTP_PROXY
-ENV HTTPS_PROXY=$HTTPS_PROXY
 RUN <<EOT
 #!/bin/bash
-curl --retry 5 https://get.modular.com | sh - 
-mamba create -n mojo python=3.10 ipykernel ipython nglview -c conda-forge -y 
-mamba activate mojo && modular install mojo
-mamba run -n mojo python -m ipykernel install --user --name="mojo" --display-name="mojo_env"
+curl https://get.modular.com | sh - 
+mamba create -n mojo python=3.10 -c conda-forge -y 
+mamba activate mojo
+modular install mojo
 EOT
 
-# Install Rust https://rsproxy.cn/#getStarted
-ENV RUSTUP_DIST_SERVER="https://rsproxy.cn"
-ENV RUSTUP_UPDATE_ROOT="https://rsproxy.cn/rustup"
+# Install Rust
 ENV PATH="/home/${CREATE_USER}/.cargo/bin:$PATH"
-ENV CARGO_UNSTABLE_SPARSE_REGISTRY="true"
 RUN <<EOT
 #!/bin/bash
-curl --proto '=https' --tlsv1.2 -sSf https://rsproxy.cn/rustup-init.sh | sh -s -- -y
-echo '[source.crates-io]
-replace-with = "rsproxy-sparse"
-[source.rsproxy]
-registry = "https://rsproxy.cn/crates.io-index"
-[source.rsproxy-sparse]
-registry = "sparse+https://rsproxy.cn/index/"
-[registries.rsproxy]
-index = "https://rsproxy.cn/crates.io-index"
-[net]
-git-fetch-with-cli = true' >> ~/.cargo/config.toml
+curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh -s -- -y
 source $HOME/.cargo/env
 cargo install evcxr_jupyter
 evcxr_jupyter --install
 EOT
 
-RUN <<EOT
-#!/bin/bash
-sudo apt update
-sudo apt install -y texlive-full
-mamba create -n torch tensorboard jupyter_packaging pytorch biopython biopandas pymol-open-source mdtraj torchvision torchaudio python=3 ipython requests scienceplots autopep8 ipykernel\
- pandas numpy matplotlib scipy seaborn orange3 -c pytorch -c nvidia -c conda-forge -y 
-git clone https://ghproxy.dockless.eu.org/https://github.com/arose/nglview
-cd nglview
-mamba run -n torch python3 setup.py install
-cd js
-rm -rf node_modules package-lock.json
-npm install typescript@latest --registry=https://registry.npmmirror.com
-npm install --registry=https://registry.npmmirror.com
-cd ../..
-mamba run -n torch python -m ipykernel install --user --name="torch" --display-name="torch_env"
-EOT
-
 WORKDIR "${HOME}"
 

spawnerdockerfile/Dockerfile.ngc

@@ -15,6 +15,7 @@ ENV ROOT_PASSWD=${ROOT_PASSWD}
 WORKDIR /root
 SHELL ["/bin/bash", "-c"]
 
+ENV STAGE_DIR=/tmp
 # https://network.nvidia.com/products/infiniband-drivers/linux/mlnx_ofed/
 ENV MLNX_OFED_VERSION=23.10-3.2.2.0 
 RUN <<EOT
@@ -29,34 +30,53 @@ python3 -m pip uninstall -y deepspeed
 # from https://github.com/huggingface/transformers/blob/main/docker/transformers-pytorch-deepspeed-latest-gpu/Dockerfile install deepspeed fail 
 # reference deepspeed install from https://github.com/microsoft/DeepSpeed/blob/master/docker/Dockerfile
 # install deepspeed prepare
+apt-get update
+apt-get install -y libnvidia-compute-535
+apt-get install -y flex tk ethtool libpci3 libltdl-dev bison lsof tcl libelf1 pciutils kmod libmnl0 debhelper libusb-1.0-0 graphviz chrpath swig libfuse2 udev
 # install Mellanox OFED
 mkdir -p ${STAGE_DIR}
+cd ${STAGE_DIR}
 wget -q -O - http://www.mellanox.com/downloads/ofed/MLNX_OFED-${MLNX_OFED_VERSION}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64.tgz | tar xzf -
-cd MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64
-./mlnxofedinstall --user-space-only --without-fw-update --all -q 
+cd ${STAGE_DIR}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64
+./mlnxofedinstall --user-space-only --without-fw-update --all -q > /tmp/mlnxofedinstall.log 2>&1
 cd ${STAGE_DIR} 
-rm -rf ${STAGE_DIR}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64*
+# rm -rf ${STAGE_DIR}/MLNX_OFED_LINUX-${MLNX_OFED_VERSION}-ubuntu22.04-x86_64*
 EOT
 
 ARG NV_PEER_MEM_VERSION="1.2"
 ENV NV_PEER_MEM_VERSION=${NV_PEER_MEM_VERSION}
 ENV NV_PEER_MEM_TAG=${NV_PEER_MEM_VERSION}-0
-RUN <<EOT
-#!/bin/bash
-# install nv_peer_mem
-mkdir -p ${STAGE_DIR}
-git clone https://github.com/Mellanox/nv_peer_memory.git --branch ${NV_PEER_MEM_TAG} ${STAGE_DIR}/nv_peer_memory
-cd ${STAGE_DIR}/nv_peer_memory
-./build_module.sh
-cd ${STAGE_DIR}
-tar xzf ${STAGE_DIR}/nvidia-peer-memory_${NV_PEER_MEM_VERSION}.orig.tar.gz
-cd ${STAGE_DIR}/nvidia-peer-memory-${NV_PEER_MEM_VERSION}
-apt-get update
-apt --fix-broken install -y
-apt-get install -y dkms
-dpkg-buildpackage -us -uc
-dpkg -i ${STAGE_DIR}/nvidia-peer-memory_${NV_PEER_MEM_TAG}_all.deb
-EOT
+RUN mkdir -p ${STAGE_DIR} && \
+    git clone https://github.com/Mellanox/nv_peer_memory.git --branch ${NV_PEER_MEM_TAG} ${STAGE_DIR}/nv_peer_memory && \
+    cd ${STAGE_DIR}/nv_peer_memory && \
+    ./build_module.sh && \
+    cd ${STAGE_DIR} && \
+    tar xzf ${STAGE_DIR}/nvidia-peer-memory_${NV_PEER_MEM_VERSION}.orig.tar.gz && \
+    cd ${STAGE_DIR}/nvidia-peer-memory-${NV_PEER_MEM_VERSION} && \
+    apt-get update && \
+    apt-get --fix-broken install -y && \
+    apt-get install -y dkms && \
+    dpkg-buildpackage -us -uc && \
+    dpkg -i ${STAGE_DIR}/nvidia-peer-memory_${NV_PEER_MEM_TAG}_all.deb
+
+# ENV NV_PEER_MEM_VERSION=${NV_PEER_MEM_VERSION}
+# ENV NV_PEER_MEM_TAG=${NV_PEER_MEM_VERSION}-0
+# RUN <<EOT
+# #!/bin/bash
+# # install nv_peer_mem
+# mkdir -p ${STAGE_DIR}
+# git clone https://github.com/Mellanox/nv_peer_memory.git --branch ${NV_PEER_MEM_TAG} ${STAGE_DIR}/nv_peer_memory
+# cd ${STAGE_DIR}/nv_peer_memory
+# ./build_module.sh
+# cd ${STAGE_DIR}
+# tar xzf ${STAGE_DIR}/nvidia-peer-memory_${NV_PEER_MEM_VERSION}.orig.tar.gz
+# cd ${STAGE_DIR}/nvidia-peer-memory-${NV_PEER_MEM_VERSION}
+# apt-get update
+# apt-get --fix-broken install -y
+# apt-get install -y dkms
+# dpkg-buildpackage -us -uc
+# dpkg -i ${STAGE_DIR}/nvidia-peer-memory_${NV_PEER_MEM_TAG}_all.deb
+# EOT
 
 # base tools
 RUN <<EOT
@@ -86,9 +106,9 @@ pip install -v -U git+https://github.com/facebookresearch/xformers.git@main#egg=
 pip install git+https://github.com/huggingface/transformers
 EOT
 
-ENV STAGE_DIR=/tmp
 RUN <<EOT
 #!/bin/bash
+apt install -y cmake gcc g++ libssl-dev
 git clone https://github.com/microsoft/DeepSpeed-Kernels.git ${STAGE_DIR}/DeepSpeed-Kernels
 cd ${STAGE_DIR}/DeepSpeed-Kernels
 python -m pip install -v .
@@ -142,6 +162,9 @@ python -m pip install black~=23.1 ruff==0.1.5 diffusers>=0.17.0
 python -m pip install --no-deps git+https://github.com/huggingface/optimum.git#egg=optimum[diffusers,quality]
 EOT
 
+
+# SSH daemon port inside container cannot conflict with host OS port
+ENV SSH_PORT=2222
 RUN <<EOT
 #!/bin/bash
 # 项目目录中的定义通常会覆盖用户家目录中的定义
@@ -156,6 +179,8 @@ CUDA_HOME=${CUDA_HOME}
 LD_LIBRARY_PATH=${LD_LIBRARY_PATH}
 EOF
 unset https_proxy http_proxy
+cat /etc/ssh/sshd_config > ${STAGE_DIR}/sshd_config && \
+sed "0,/^Port 22/s//Port ${SSH_PORT}/" ${STAGE_DIR}/sshd_config > /etc/ssh/sshd_config
 EOT
 
 CMD ["/usr/sbin/sshd", "-D"]

spawnerdockerfile/README.md

@@ -29,7 +29,8 @@ NCCL_IB_DISABLE=0
 ## test command
 
 docker run -it --rm --network=host --privileged --ipc=host --ulimit memlock=-1 --gpus all hotwa/notebook:ngc
-docker run --rm -it --gpus all --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 hotwa/notebook:ngc /bin/bash
+docker run --rm -it --gpus all --ipc=host --ulimit memlock=-1 --ulimit stack=67108864 nvcr.io/nvidia/pytorch:24.06-py3 /bin/bash
+docker run -it --rm --network=host --privileged --ipc=host --ulimit memlock=-1 --gpus all quay.io/hotwa/ngc:latest /bin/bash
 
 ```shell
 nvidia-smi

spawnerdockerfile/docker-compose_ngc.yml

@@ -45,7 +45,7 @@ services:
     # tty: true
     # stdin_open: true
     restart: unless-stopped
-    image: hotwa/notebook:ngc
+    image: quay.io/hotwa/ngc:latest
     privileged: true
     ipc: host
     network_mode: host