Taskmaster

Taskmaster is a project developed as part of the 42 School curriculum. The goal of this project is to create a job control daemon, similar to supervisor, that can manage and supervise processes based on a configuration file. The program allows users to start, stop, restart, and monitor processes, as well as reload the configuration dynamically.

In our implementation, we have extended the project by introducing a client/server architecture, where the server handles the job control and the client provides an interactive shell for the user to manage the processes.

Features

• Job Control: Start, stop, restart, and monitor processes.
• Configuration File: Define processes to supervise, including command, number of processes, auto-start, restart policies, and more.
• Dynamic Configuration Reload: Reload the configuration file without stopping the main program by sending a SIGHUP signal.
• Logging System: Log events such as process starts, stops, restarts, and unexpected deaths to a local file.
• Interactive Shell: A control shell for users to interact with the job control daemon.
• Client/Server Architecture:
  • Server: Runs as a daemon and manages the processes.
  • Client: Provides an interactive shell to communicate with the server and control the processes.

Run the project

Run the server:

cargo run --bin server

Run the client:

cargo run --bin client

Configuration

The configuration file defines the programs to be supervised. Here is an example configuration file in YAML format:

  nginx:
    cmd: "ping google.com"
    numprocs: 2
    umask: 77
    workingdir: /tmp
    autostart: true
    autorestart: unexpected
    exitcodes:
      - 0
      - 2
    startretries: 3
    starttime: 5
    stopsignal: SIGTERM
    stoptime: 5
    stdout: /tmp/nginx.stdout
    stderr: /tmp/nginx.stderr
    env:
      STARTED_BY: taskmaster
      ANSWER: 42

  vogsphere:
    cmd: "/bin/ls"
    numprocs: 2
    umask: 77
    workingdir: /tmp
    autostart: true
    autorestart: unexpected
    exitcodes:
      - 0
      - 2
    startretries: 3
    starttime: 5
    stopsignal: SIGUSR1
    stoptime: 10
    stdout: /tmp/vgsworker.stdout
    stderr: /tmp/vgsworker.stderrprograms
    fatal_state_report_address: "127.0.0.1:8080"

Monitoring State Machine

stateDiagram
    [*] --> NotStartedYet : Config
    NotStartedYet --> Starting : Config or User
    Starting --> Running : Process Listening (use config setting to determine how much time should be left to the program to consider it started)
    Starting --> Backoff : Process Listening (use config setting to determine when enough retried have been done)
    Running --> Stopping : User (use config setting to determine which signal to send)
    Running --> ExitedExpectedly : Process Listening (use config setting to determine if the received exit code is consider an Expected exit)
    Running --> ExitedUnExpectedly : Process Listening (use config setting to determine if the received exit code is consider a unExpected exit)
    Backoff --> Starting : Config
    Backoff --> Fatal : Config
    Stopping --> Stopped : Process Listening (use config setting to determine when to kill the program)
    ExitedExpectedly --> Starting : Config or User
    ExitedUnExpectedly --> Starting : Config or User
    Fatal --> Starting : User
    Stopped --> Starting : User

Loading