Improved Installation / Prepare Script (#160)

* Slimmed down prepare * Revert "New syntax for swarm" This reverts commit fa942ce. * Update libnvidia-container * Avoid prompts * Sequence * Syntax harmonize * WSL2 compatible detection * Case insensitive Microsoft * Remove set -e * Tune final message
aws-deepracer-community · Apr 7, 2024 · 2daf796 · 2daf796
1 parent 10b8483
commit 2daf796
Showing 1 changed file with 62 additions and 58 deletions.
diff --git a/bin/prepare.sh b/bin/prepare.sh
@@ -1,4 +1,6 @@
-#!/bin/bash
+#!/usr/bin/env bash
+
+set -e
 
 trap ctrl_c INT
 
@@ -7,81 +9,75 @@ function ctrl_c() {
     exit 1
 }
 
+export DEBIAN_FRONTEND=noninteractive
 DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" >/dev/null 2>&1 && pwd)"
 
+## Check distribution
+distribution=$(
+    . /etc/os-release
+    echo $ID$VERSION_ID | sed 's/\.//'
+)
+
+## Check if WSL2
+if grep -qi Microsoft /proc/version && grep -q "WSL2" /proc/version; then
+    IS_WSL2="yes"
+fi
+
+## Remove needsreboot in Ubuntu 22.04
+if [[ "${distribution}" == "ubuntu2204" && -z "${IS_WSL2}" ]]; then
+    sudo apt remove -y needrestart
+fi
+
 ## Patch system
-sudo apt-get update && sudo apt-mark hold grub-pc && sudo DEBIAN_FRONTEND=noninteractive apt-get -y -o \
+sudo apt update && sudo apt-mark hold grub-pc && sudo apt -y -o \
     DPkg::options::="--force-confdef" -o DPkg::options::="--force-confold" -qq --force-yes upgrade &&
-    sudo apt-get install --no-install-recommends -y jq
+    sudo apt install --no-install-recommends -y jq awscli python3-boto3
 source $DIR/detect.sh
 echo "Detected cloud type ${CLOUD_NAME}"
 
 ## Do I have a GPU
-GPUS=$(lspci | awk '/NVIDIA/ && ( /VGA/ || /3D controller/ ) ' | wc -l)
+GPUS=0
+if [[ -z "${IS_WSL2}" ]]; then
+    GPUS=$(lspci | awk '/NVIDIA/ && ( /VGA/ || /3D controller/ ) ' | wc -l)
+else
+    if [[ -f /usr/lib/wsl/lib/nvidia-smi ]]; then
+        GPUS=$(nvidia-smi --query-gpu=name --format=csv,noheader | wc -l)
+    fi
+fi
 if [ $? -ne 0 ] || [ $GPUS -eq 0 ]; then
     ARCH="cpu"
     echo "No NVIDIA GPU detected. Will not install drivers."
 else
     ARCH="gpu"
 fi
 
-## Do I have an additional disk for Docker images - looking for /dev/sdc (Azure)
-
-if [[ "${CLOUD_NAME}" == "azure" ]]; then
-    ADDL_DISK=$(lsblk | awk '/^sdc/ {print $1}')
-    ADDL_PART=$(lsblk -l | awk -v DISK="$ADDL_DISK" '($0 ~ DISK) && ($0 ~ /part/) {print $1}')
-
-    if [ -n "$ADDL_DISK" ] && [ -z "$ADDL_PART" ]; then
-        echo "Found $ADDL_DISK, preparing it for use"
-        echo -e "g\nn\np\n1\n\n\nw\n" | sudo fdisk /dev/$ADDL_DISK
-        sleep 1s
-        ADDL_DEVICE=$(echo "/dev/"$ADDL_DISK"1")
-        sudo mkfs.ext4 $ADDL_DEVICE
-        sudo mkdir -p /var/lib/docker
-        echo "$ADDL_DEVICE   /var/lib/docker   ext4    rw,user,auto    0    0" | sudo tee -a /etc/fstab
-        mount /var/lib/docker
-        if [ $? -ne 0 ]; then
-            echo "Error during preparing of additional disk. Exiting."
-            exit 1
-        fi
-    elif [ -n "$ADDL_DISK" ] && [ -n "$ADDL_PART" ]; then
-        echo "Found $ADDL_DISK - $ADDL_PART already mounted. Installing into present drive/directory structure."
-
-    else
-        echo "Did not find $ADDL_DISK. Installing into present drive/directory structure."
-    fi
-fi
-
 ## Adding Nvidia Drivers
-if [[ "${ARCH}" == "gpu" ]]; then
-    distribution=$(
-        . /etc/os-release
-        echo $ID$VERSION_ID | sed 's/\.//'
-    )
-    curl -sS https://developer.download.nvidia.com/compute/cuda/repos/$distribution/x86_64/3bf863cc.pub | sudo gpg --dearmour -o /etc/apt/trusted.gpg.d/nvidia-cuda.gpg
-    curl -sS https://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu2004/x86_64/7fa2af80.pub | sudo gpg --dearmour -o /etc/apt/trusted.gpg.d/nvidia-machine-learning.gpg
-    echo "deb http://developer.download.nvidia.com/compute/cuda/repos/$distribution/x86_64 /" | sudo tee /etc/apt/sources.list.d/cuda.list
-    echo "deb http://developer.download.nvidia.com/compute/machine-learning/repos/ubuntu2004/x86_64 /" | sudo tee /etc/apt/sources.list.d/cuda_learn.list
-    sudo apt update && sudo apt install -y nvidia-driver-470-server cuda-minimal-build-11-4 --no-install-recommends -o Dpkg::Options::="--force-overwrite"
+if [[ "${ARCH}" == "gpu" && -z "${IS_WSL2}" ]]; then
+    case $distribution in
+    ubuntu2004)
+        sudo apt install -y nvidia-driver-525-server --no-install-recommends -o Dpkg::Options::="--force-overwrite"
+        ;;
+    ubuntu2204)
+        sudo apt install -y nvidia-driver-535-server --no-install-recommends -o Dpkg::Options::="--force-overwrite"
+        ;;
+    *)
+        echo "Unsupported distribution: $distribution"
+        exit 1
+        ;;
+    esac
 fi
 
-## Adding AWSCli
-sudo apt-get install -y --no-install-recommends awscli python3-boto3
-
 ## Installing Docker
-curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /etc/apt/trusted.gpg.d/docker.gpg
-sudo add-apt-repository "deb [arch=amd64] https://download.docker.com/linux/ubuntu $(lsb_release -cs) stable"
-sudo apt-get update && sudo apt-get install -y --no-install-recommends docker-ce docker-ce-cli docker-buildx-plugin docker-compose-plugin containerd.io
+sudo apt install -y --no-install-recommends docker.io docker-buildx docker-compose-v2
 
+## Install Nvidia Docker Container
 if [[ "${ARCH}" == "gpu" ]]; then
-    distribution=$(
-        . /etc/os-release
-        echo $ID$VERSION_ID
-    )
-    curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo gpg --dearmor -o /etc/apt/trusted.gpg.d/nvidia-docker.gpg
-    curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
-
-    sudo apt-get update && sudo apt-get install -y --no-install-recommends nvidia-docker2 nvidia-container-toolkit nvidia-container-runtime
+    curl -fsSL https://nvidia.github.io/libnvidia-container/gpgkey | sudo gpg --dearmor -o /usr/share/keyrings/nvidia-container-toolkit-keyring.gpg &&
+        curl -s -L https://nvidia.github.io/libnvidia-container/stable/deb/nvidia-container-toolkit.list |
+        sed 's#deb https://#deb [signed-by=/usr/share/keyrings/nvidia-container-toolkit-keyring.gpg] https://#g' |
+            sudo tee /etc/apt/sources.list.d/nvidia-container-toolkit.list
+
+    sudo apt update && sudo apt install -y --no-install-recommends nvidia-docker2 nvidia-container-runtime
     if [ -f "/etc/docker/daemon.json" ]; then
         echo "Altering /etc/docker/daemon.json with default-rutime nvidia."
         cat /etc/docker/daemon.json | jq 'del(."default-runtime") + {"default-runtime": "nvidia"}' | sudo tee /etc/docker/daemon.json
@@ -90,21 +86,29 @@ if [[ "${ARCH}" == "gpu" ]]; then
         sudo cp $DIR/../defaults/docker-daemon.json /etc/docker/daemon.json
     fi
 fi
-sudo systemctl enable docker
-sudo systemctl restart docker
+
+## Enable and start docker
+if [[ -n "${IS_WSL2}" ]]; then
+    sudo service docker restart
+else
+    sudo systemctl enable docker
+    sudo systemctl restart docker
+fi
 
 ## Ensure user can run docker
 sudo usermod -a -G docker $(id -un)
 
 ## Reboot to load driver -- continue install if in cloud-init
 CLOUD_INIT=$(pstree -s $BASHPID | awk /cloud-init/ | wc -l)
 
-if [[ "$CLOUD_INIT" -ne 0 ]]; then
+if [[ "${CLOUD_INIT}" -ne 0 ]]; then
     echo "Rebooting in 5 seconds. Will continue with install."
     cd $DIR
     ./runonce.sh "./init.sh -c ${CLOUD_NAME} -a ${ARCH}"
     sleep 5s
     sudo shutdown -r +1
+elif [[ -n "${IS_WSL2}" || "${ARCH}" == "cpu" ]]; then
+    echo "First stage done. Log out, then log back in and run init.sh -c ${CLOUD_NAME} -a ${ARCH}"
 else
     echo "First stage done. Please reboot and run init.sh -c ${CLOUD_NAME} -a ${ARCH}"
 fi