Setup

Before performing any deployment task, run the Setup tasks below. Once done, run the Cleanup tasks below. If you haven’t already, please follow the Get started guide.

1. Update deploy repositories

Ensure the deploy, pillar/private and salt/private repositories are on the default branch and are up-to-date. You can run this convenience script to run the appropriate git commands:

./script/update

Check the output in case there are any issues switching to the default branch or any conflicts pulling from GitHub.

2. Check if Kingfisher is busy

Note

Skip this step unless you’re working on Kingfisher.

  1. Access Scrapyd’s web interface, click Jobs and look under Running. If any spiders are running, don’t deploy without the consent of data support managers.

  2. SSH into collect.kingfisher.open-contracting.org as the root user.

  3. Check if any long-running tasks are running, by attaching to each session in tmux to see which commands are running. If any commands would be interrupted by the deployment, don’t deploy without the consent of the data support managers, who should be identified by the session names.

    To list all sessions:

    for i in root $(ls -1 /home); do echo $i; su $i -c "tmux ls"; done

  4. If the postgres service would be restarted by the deployment (for example, due to a configuration change or a package upgrade), check if any long-running queries are running. If there are queries with a state of active and a time greater than an hour, don’t deploy without the consent of the data support managers, who should be identified by the usename, client_addr or comment at the start of query.