Skip to content

USER GUIDE

Hook: carrier / no-carrier

Purpose

Run scripts when the link layer goes UP (carrier.d) or DOWN (no-carrier.d) — cable plug, Wi‑Fi associate, or switch port flap — before L3 addresses exist.

When to use it

  • Notify ops when a physical link returns (PDU cycle, switch maintenance)
  • Trigger low-level diagnostics that do not need an IP yet
  • Not for “has an IP / default route” — use Hook: routable for L3

How to get there

  • Directories: /etc/netevd/carrier.d/, /etc/netevd/no-carrier.d/
  • Backends: all (systemd-networkd, NetworkManager, dhclient)
  • Scripts must be executable; run in lexical order (10-, 20-, …)

Operate from CLI

  1. Install a carrier-up logger:
sudo tee /etc/netevd/carrier.d/10-link-up.sh >/dev/null <<'EOF'
#!/bin/bash
logger -t netevd "carrier UP on $LINK (index $LINKINDEX, backend $BACKEND)"
EOF
sudo chmod +x /etc/netevd/carrier.d/10-link-up.sh
  1. Install a no-carrier handler:
sudo tee /etc/netevd/no-carrier.d/10-link-down.sh >/dev/null <<'EOF'
#!/bin/bash
logger -t netevd "carrier DOWN on $LINK"
EOF
sudo chmod +x /etc/netevd/no-carrier.d/10-link-down.sh
  1. Validate config and bounce the link to test:
netevd validate
sudo ip link set eth0 down && sudo ip link set eth0 up
netevd events -f -i eth0 -t carrier &
journalctl -u netevd -f
  1. Dry-run hook env without executing:
netevd test /etc/netevd/carrier.d/10-link-up.sh -i eth0 -t carrier

  1. Empty / fail: Script never runs → missing +x, wrong directory, or YAML filter with action: ignore on docker*/pattern match; check journalctl -u netevd for exit codes (non-zero exits are logged, other scripts still run).

  2. Success: logger or script output appears on link up/down; $LINK and $BACKEND are set; carrier events visible in netevd events.