Skip to content

Latest commit

 

History

History
executable file
·
102 lines (85 loc) · 4.49 KB

README.md

File metadata and controls

executable file
·
102 lines (85 loc) · 4.49 KB

Automation Scripts

This repository contains automation scripts for setting up a swarm learning environment on a cluster of machines.

Contents

This script automates the process of setting up a workspace for swarm learning. It puts together several other scripts to perform the following tasks:

  • Check for open and exposed ports on the host server
  • Install prerequisites for the workspace
  • Install necessary containers for the workspace
  • Set up GPU environment for the workspace
  • Generate and share SSL certificate for the workspace
  • Set up swarm learning environment
  • Set up swarm network node and/or sentinel node
  • Share the SSL certificate with other nodes in the cluster
  • Configure swarm learning environment for the workspace

Usage

To use the scripts, clone this repository onto your machine and run automate.sh with the required flags:

$ sh workspace/automate_scripts/automate.sh [-a|-b|-c] [-w workspace_name] [-d host_index] [-s sentinel_ip] [-n num_peers] [-e num_epochs] [-h]"

Help

To see the usage instructions, run:

$ sh workspace/automate_scripts/automate.sh -h

The flags are as follows:

  • [-a|-b|-c] indicates the build stage. The script has three build stages and please ensure that you run the scripts in the correct order. The three stages are:
    • Prerequisite: Runs scripts that check for required software and open/exposed ports.
    • Server setup: Runs scripts that set up the swarm learning environment on a server.
    • Final sl setup: Downloads the dataset for the workspace. This will take a long time. The [-s sentinel_ip] flag is necessary. The script will download the dataset from the sentinel node.
  • -w: Name of the workspace module, e.g. odelia-breast-mri, marugoto_mri, etc.
  • -d: Host index, host index should be chosen from [TUD, Ribera, VHIO, Radboud, UKA, Utrecht, Mitera, Cambridge, Zurich]
  • -s: Sentinel IP address, e.g. 172.24.40.65
  • -p: Minimum number of peers involved in the next training tast for swarm learning. We should coordinate this with the number of peers in the cluster.
  • -e: Number of epochs for swarm learning. Optional. Default is 50 for marugoto_mri and 100 for odelia-breast-mri.

The script has three main actions:

  • Prerequisite: Runs scripts that check for required software and open/exposed ports.
$ sh workspace/automate_scripts/automate.sh -a
  • Server setup: Runs scripts that set up the swarm learning environment on a server.
$ sh workspace/automate_scripts/automate.sh -b -s <sentinel_ip> -d <host_index>
  • Final sl setup: Runs scripts that finalize the setup of the swarm learning environment. Only <> is required. The [-n num_peers] and [-e num_epochs] flags are optional.
$ sh workspace/automate_scripts/automate.sh -c -w <workspace_name> -s <sentinel_ip> -d <host_index> [-n num_peers] [-e num_epochs]

Here is an example usage for flag -c:

$ sh workspace/automate_scripts/automate.sh -c -w odelia-breast-mri -s 172.24.40.65 -d VHIO -n 3 -e 50

This will run the prerequisite, server_setup, and final_setup actions with the specified parameters.

Downloading the preprocessed dataset

  • Download dataset: Downloads the dataset for the workspace. This will take a long time, could run it at any time. Just run with either command, get_dataset_gdown.sh without vpn, or get_dataset_scp.sh with vpn. get_dataset_gdown.sh will download the dataset from google drive.
$ sh workspace/automate_scripts/sl_env_setup/get_dataset_gdown.sh

The [-s sentinel_ip] flag is only necessary for get_dataset_scp.sh The script will download the dataset from the sentinel node.

$ sh workspace/automate_scripts/sl_env_setup/get_dataset_scp.sh -s <sentinel_ip>

VPN setup

  • The setup of VPN is included in the automate.sh script. Just ask @Jeff for credentials when the terminal asks for it.
  • Optional: To configure OpenVPN to only take effect on the 'swarm' user and not affect others, you can follow the steps below:
    1. Create a new group for 'swarm' user
    $ sudo groupadd swarm-group
    1. Add the 'swarm' user to the new group
    $ sudo usermod -a -G swarm-group swarm
    1. Modify the OpenVPN server configuration file
    $ sudo nano /etc/openvpn/server.conf
    1. Add the following lines to the configuration file to specify the user and group to run OpenVPN user swarm
    $ group swarm-group
    1. Save the configuration file and restart OpenVPN
    $ sudo systemctl restart openvpn