Network problems hit you out of context — a service is unreachable, a script can't talk to a database, the dashboard says "host down". Don't guess. RHEL gives you a clean four-layer triage that walks the OSI stack in the order most failures actually occur:
- Link — is the interface up, does it have a MAC? (
ip link, nmcli device)
- Address — does it have an IP, the right one? (
ip address, nmcli connection show)
- Routing — can the kernel reach the destination's network? (
ip route, ping <gateway>)
- Resolution — can it turn names into addresses? (
getent hosts, cat /etc/resolv.conf)
A failure at any layer makes higher layers look broken. If your IP is wrong (Layer 2), every name-resolution test will also time out — but fixing the IP fixes both. Always start at the bottom.
The fix layer is almost always NetworkManager — nmcli connection modify to repair, systemctl restart NetworkManager to bounce the daemon if connections are stuck.
Cert