Skip to content

README.md

Latest commit

 

History

History
91 lines (62 loc) · 2.71 KB

README.md

File metadata and controls

91 lines (62 loc) · 2.71 KB

SYNC.TOOLING

CI Docs Coverage Python Version

Tools for monitoring and diagnosing time synchronization in a distributed system.

The tools currently handle monitoring

  • PTP4L (port states, clock state)
  • PHC2SYS (servo state, offset)
  • local and remote PTP instances by implementing a PTP Management Client

Documentation is available at tier4.github.io/sync_tooling 📖

System Requirements

  • Ubuntu >= 22.04
  • ROS 2 >= Humble
  • Python >= 3.10

For diag-worker, the following additional requirements apply:

  • Ethtool (sudo apt install ethtool)
  • PTP4L and PHC2SYS (sudo apt install linuxptp)

Installation

SYNC.TOOLING has to be installed on every machine that will run a diag-master or a diag-worker.

Prerequisites

This project uses uv as its package manager. You can install it via

pip install uv

Development Build

This method builds the project locally for development purposes.

git clone --recursive https://github.com/tier4/sync_tooling.git
cd sync_tooling

uv sync --all-packages

# Replace `humble` with your ROS 2 distro name, e.g. jazzy.
source /opt/ros/humble/setup.bash
uv run pytest

Production Build

This method builds Python packages (wheel files) that can be installed using uv or pip.

git clone --recursive https://github.com/tier4/sync_tooling.git
cd sync_tooling

uv build --all-packages

This will generate a dist directory with the built packages, and the dist/*.whl files can be installed using pip:

pip install dist/*.whl

Usage

# In the `sync_tooling` folder checked out above
uv run diag-master --reference diag_master/config/sample.yml

# The diag-worker needs to run privileged so that it can communicate with local Unix domain 
# sockets and read the journal of services running with elevated privileges
# See notes below.
sudo uv run diag-worker --ptp4l-units ptp4l@eno1 ptp4l@enp2s0

Note: To run uv privileged, it needs to be installed system-wide. If sudo uv does not work after a normal pip install uv, try sudo pip install uv.