[Image] Fix blocking by unattended-upgrade (skypilot-org#1347)

* Fix blocking by unattended-upgrade

* adopt to gcp and azure

Author: Michaelvll

sky/templates/aws-ray.yml.j2

@@ -89,6 +89,7 @@ initialization_commands: []
 # Increment the following for catching performance bugs easier:
 #   current num items (num SSH connections): 1
 setup_commands:
+  # Disable `unattended-upgrades` to prevent apt-get from hanging. It should be called at the beginning before the process started to avoid being blocked. (This is a temporary fix.)
   # Create ~/.ssh/config file in case the file does not exist in the custom image.
   # Make sure python3 & pip3 are available on this image.
   # We set auto_activate_base to be false for pre-installed conda.
@@ -97,19 +98,21 @@ setup_commands:
   # Line 'sudo grep ..': set the number of threads per process to unlimited to avoid ray job submit stucking issue when the number of running ray jobs increase.
   # Line 'mkdir -p ..': disable host key check
   # Line 'python3 -c ..': patch the buggy ray files and enable `-o allow_other` option for `goofys`
-  - mkdir -p ~/.ssh; touch ~/.ssh/config;
+  - sudo systemctl stop unattended-upgrades || true;
+    sudo systemctl disable unattended-upgrades || true;
+    sudo sed -i 's/Unattended-Upgrade "1"/Unattended-Upgrade "0"/g' /etc/apt/apt.conf.d/20auto-upgrades || true;
+    sudo kill -9 `sudo lsof /var/lib/dpkg/lock-frontend | awk '{print $2}' | tail -n 1` || true;
+    sudo pkill -9 apt-get;
+    sudo pkill -9 dpkg;
+    sudo dpkg --configure -a;
+    mkdir -p ~/.ssh; touch ~/.ssh/config;
     pip3 --version > /dev/null 2>&1 || (curl -sSL https://bootstrap.pypa.io/get-pip.py -o get-pip.py && python3 get-pip.py && echo "PATH=$HOME/.local/bin:$PATH" >> ~/.bashrc);
     (type -a python | grep -q python3) || echo 'alias python=python3' >> ~/.bashrc;
     (type -a pip | grep -q pip3) || echo 'alias pip=pip3' >> ~/.bashrc;
     (which conda > /dev/null 2>&1 && conda init > /dev/null && conda config --set auto_activate_base false) || (wget -nc https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh && bash Miniconda3-latest-Linux-x86_64.sh -b && eval "$(~/miniconda3/bin/conda shell.bash hook)" && conda init && conda config --set auto_activate_base true);
     source ~/.bashrc;
     (pip3 list | grep ray | grep {{ray_version}} 2>&1 > /dev/null || pip3 install -U ray[default]=={{ray_version}}) && mkdir -p ~/sky_workdir && mkdir -p ~/.sky/sky_app;
     (pip3 list | grep skypilot && [ "$(cat {{sky_remote_path}}/current_sky_wheel_hash)" == "{{sky_wheel_hash}}" ]) || (pip3 uninstall skypilot -y; pip3 install "$(echo {{sky_remote_path}}/{{sky_wheel_hash}}/skypilot-{{sky_version}}*.whl)[aws]" && echo "{{sky_wheel_hash}}" > {{sky_remote_path}}/current_sky_wheel_hash || exit 1);
-    sudo systemctl stop unattended-upgrades;
-    sudo kill -9 `sudo lsof /var/lib/dpkg/lock-frontend | awk '{print $2}' | tail -n 1` || true;
-    sudo pkill -9 apt-get;
-    sudo pkill -9 dpkg;
-    sudo dpkg --configure -a;
     sudo bash -c 'rm -rf /etc/security/limits.d; echo "* soft nofile 1048576" >> /etc/security/limits.conf; echo "* hard nofile 1048576" >> /etc/security/limits.conf';
     sudo grep -e '^DefaultTasksMax' /etc/systemd/system.conf || (sudo bash -c 'echo "DefaultTasksMax=infinity" >> /etc/systemd/system.conf'); sudo systemctl set-property user-$(id -u $(whoami)).slice TasksMax=infinity; sudo systemctl daemon-reload;
     mkdir -p ~/.ssh; (grep -Pzo -q "Host \*\n  StrictHostKeyChecking no" ~/.ssh/config) || printf "Host *\n  StrictHostKeyChecking no\n" >> ~/.ssh/config;

sky/templates/azure-ray.yml.j2

@@ -88,26 +88,29 @@ initialization_commands: []
 # Increment the following for catching performance bugs easier:
 #   current num items (num SSH connections): 1
 setup_commands:
+  # Disable `unattended-upgrades` to prevent apt-get from hanging. It should be called at the beginning before the process started to avoid being blocked. (This is a temporary fix.)
   # Create ~/.ssh/config file in case the file does not exist in the image.
   # Make sure python3 & pip3 are available on this image.
   # Line 'sudo bash ..': set the ulimit as suggested by ray docs for performance. https://docs.ray.io/en/latest/cluster/vms/user-guides/large-cluster-best-practices.html#system-configuration
   # Line 'sudo grep ..': set the number of threads per process to unlimited to avoid ray job submit stucking issue when the number of running ray jobs increase.
   # Line 'mkdir -p ..': disable host key check
   # Line 'python3 -c ..': patch the buggy ray files and enable `-o allow_other` option for `goofys`
   # This also kills the service that is holding the lock on dpkg (problem only exists on aws/azure, not gcp)
-  - mkdir -p ~/.ssh; touch ~/.ssh/config;
+  - sudo systemctl stop unattended-upgrades || true;
+    sudo systemctl disable unattended-upgrades || true;
+    sudo sed -i 's/Unattended-Upgrade "1"/Unattended-Upgrade "0"/g' /etc/apt/apt.conf.d/20auto-upgrades || true;
+    sudo kill -9 `sudo lsof /var/lib/dpkg/lock-frontend | awk '{print $2}' | tail -n 1` || true;
+    sudo pkill -9 apt-get;
+    sudo pkill -9 dpkg;
+    sudo dpkg --configure -a;
+    mkdir -p ~/.ssh; touch ~/.ssh/config;
     pip3 --version > /dev/null 2>&1 || (curl -sSL https://bootstrap.pypa.io/get-pip.py -o get-pip.py && python3 get-pip.py && echo "PATH=$HOME/.local/bin:$PATH" >> ~/.bashrc);
     (type -a python | grep -q python3) || echo 'alias python=python3' >> ~/.bashrc;
     (type -a pip | grep -q pip3) || echo 'alias pip=pip3' >> ~/.bashrc;
     which conda > /dev/null 2>&1 || (wget -nc https://repo.anaconda.com/miniconda/Miniconda3-latest-Linux-x86_64.sh && bash Miniconda3-latest-Linux-x86_64.sh -b && eval "$(/home/azureuser/miniconda3/bin/conda shell.bash hook)" && conda init && conda config --set auto_activate_base true);
     source ~/.bashrc;
     (pip3 list | grep ray | grep {{ray_version}} 2>&1 > /dev/null || pip3 install -U ray[default]=={{ray_version}}) && mkdir -p ~/sky_workdir && mkdir -p ~/.sky/sky_app && touch ~/.sudo_as_admin_successful;
     (pip3 list | grep skypilot && [ "$(cat {{sky_remote_path}}/current_sky_wheel_hash)" == "{{sky_wheel_hash}}" ]) || (pip3 uninstall skypilot -y; pip3 install "$(echo {{sky_remote_path}}/{{sky_wheel_hash}}/skypilot-{{sky_version}}*.whl)[azure]" && echo "{{sky_wheel_hash}}" > {{sky_remote_path}}/current_sky_wheel_hash || exit 1);
-    sudo systemctl stop unattended-upgrades;
-    sudo kill -9 `sudo lsof /var/lib/dpkg/lock-frontend | awk '{print $2}' | tail -n 1` || true;
-    sudo pkill -9 apt-get;
-    sudo pkill -9 dpkg;
-    sudo dpkg --configure -a;
     sudo bash -c 'rm -rf /etc/security/limits.d; echo "* soft nofile 1048576" >> /etc/security/limits.conf; echo "* hard nofile 1048576" >> /etc/security/limits.conf';
     sudo grep -e '^DefaultTasksMax' /etc/systemd/system.conf || (sudo bash -c 'echo "DefaultTasksMax=infinity" >> /etc/systemd/system.conf'); sudo systemctl set-property user-$(id -u $(whoami)).slice TasksMax=infinity; sudo systemctl daemon-reload;
     mkdir -p ~/.ssh; (grep -Pzo -q "Host \*\n  StrictHostKeyChecking no" ~/.ssh/config) || printf "Host *\n  StrictHostKeyChecking no\n" >> ~/.ssh/config;

sky/templates/gcp-ray.yml.j2

@@ -123,6 +123,7 @@ initialization_commands: []
 # Increment the following for catching performance bugs easier:
 #   current num items (num SSH connections): 1  (+1 if tpu_vm)
 setup_commands:
+  # Disable `unattended-upgrades` to prevent apt-get from hanging. It should be called at the beginning before the process started to avoid being blocked. (This is a temporary fix.)
   # Line 'mkdir -p ..': Create ~/.ssh/config file in case the file does not exist in the custom image.
   # Line 'pip3 --v ..': Make sure python3 & pip3 are available on this image.
   # Line 'which conda ..': some images (TPU VM) do not install conda by
@@ -132,7 +133,14 @@ setup_commands:
   # Line 'sudo grep ..': set the number of threads per process to unlimited to avoid ray job submit stucking issue when the number of running ray jobs increase.
   # Line 'mkdir -p ..': disable host key check
   # Line 'python3 -c ..': patch the buggy ray files and enable `-o allow_other` option for `goofys`
-  - mkdir -p ~/.ssh; touch ~/.ssh/config;
+  - sudo systemctl stop unattended-upgrades || true;
+    sudo systemctl disable unattended-upgrades || true;
+    sudo sed -i 's/Unattended-Upgrade "1"/Unattended-Upgrade "0"/g' /etc/apt/apt.conf.d/20auto-upgrades || true;
+    sudo kill -9 `sudo lsof /var/lib/dpkg/lock-frontend | awk '{print $2}' | tail -n 1` || true;
+    sudo pkill -9 apt-get;
+    sudo pkill -9 dpkg;
+    sudo dpkg --configure -a;
+    mkdir -p ~/.ssh; touch ~/.ssh/config;
     pip3 --version > /dev/null 2>&1 || (curl -sSL https://bootstrap.pypa.io/get-pip.py -o get-pip.py && python3 get-pip.py && echo "PATH=$HOME/.local/bin:$PATH" >> ~/.bashrc);
     (type -a python | grep -q python3) || echo 'alias python=python3' >> ~/.bashrc;
     (type -a pip | grep -q pip3) || echo 'alias pip=pip3' >> ~/.bashrc;