Skip to content

README.md

Latest commit

 

History

History
79 lines (58 loc) · 2.89 KB

README.md

File metadata and controls

79 lines (58 loc) · 2.89 KB

GraphHopper Update Scripts

These scripts keep GraphHopper routing data up to date using systemd timers and services.

Scripts

  • gh-update.sh — downloads OSM data, imports it into the inactive instance, switches nginx to it, stops the old instance.
  • gh-notify.sh — sends a Mailgun email on success (only when an update was actually performed) or failure.

Systemd Units

Unit Purpose
graphhopper@.service GraphHopper routing server (instances a and b)
gh-update.service Oneshot update job
gh-update.timer Triggers gh-update.service hourly
gh-update-notify@.service Email notification, called by OnSuccess=/OnFailure=

What Happens

  1. Timer triggers gh-update.service every hour.
  2. Script checks if new OSM data is available; exits silently if not.
  3. If yes, imports data into the inactive instance (a or b).
  4. Starts the new instance via systemd and polls until it is healthy.
  5. Switches nginx to the new instance and stops the old one.
  6. On success: gh-update-notify@success.service sends a notification email.
  7. On failure: gh-update-notify@failure.service sends a failure email. The service is left in failed state — the timer does not retry until the service is explicitly reset.

Setup

1. Config file

cp gh-update.conf.example gh-update.conf
chmod 600 gh-update.conf
# fill in MAILGUN_API_KEY, MAILGUN_DOMAIN, NOTIFY_EMAIL

2. Sudoers

freemap ALL=(root) NOPASSWD: /bin/systemctl reload nginx, \
  /bin/systemctl enable --now graphhopper@a, /bin/systemctl enable --now graphhopper@b, \
  /bin/systemctl disable --now graphhopper@a, /bin/systemctl disable --now graphhopper@b, \
  /bin/systemctl stop gh-update.timer, /bin/systemctl start gh-update.timer

3. Install and enable units

cp graphhopper@.service gh-update.service gh-update.timer gh-update-notify@.service \
  /etc/systemd/system/
chmod +x gh-update.sh gh-notify.sh
systemctl daemon-reload

# Start whichever graphhopper instance is currently active (e.g. a).
# The update script will enable/disable instances automatically on each update.
systemctl enable --now graphhopper@a.service

# Enable the timer (replaces cron)
systemctl enable --now gh-update.timer

Logs

journalctl -u gh-update.service       # update job
journalctl -u graphhopper@a.service   # instance a
journalctl -u graphhopper@b.service   # instance b

After a Failure

On failure the timer is stopped automatically, preventing retries until the issue is fixed. After fixing:

systemctl start gh-update.timer     # re-enable hourly runs
systemctl start gh-update.service   # optional: trigger immediately