CertGrid CertGrid
Hands-on Lab·CompTIA Linux+

Network faults, from the client outwards

Six rungs, each with one command and one unambiguous signal: the interface, the routing decision, the neighbour, the name, the port and the protocol. The centre of the page blocks ICMP on the far host and shows ping failing completely while SSH answers and the ARP entry stays reachable - three layers measured in the same second.

Troubleshooting Guide 26 of 28 Intermediate

Two hosts, because a network fault needs two ends. The only change made is a runtime ICMP block on the RPM host, added and removed within a single step.
Server NameIP AddressOSRolesCPURAMHDD
LPLUS-A01192.168.0.73Ubuntu 26.04 LTSDebian-family host - apt, ufw, netplan, AppArmor2 Core4 GB50 GB
LPLUS-B01192.168.0.74AlmaLinux 10.2RPM-family host - dnf, firewalld, NetworkManager, SELinux2 Core4 GB50 GB

This guide includes

Use this when a connection fails and you do not yet know which layer is at fault. This matters because the instinct is to start at DNS or the firewall, and because the ping that everyone runs first cannot distinguish a host that is down from one that simply does not answer ICMP.

Before you start

  1. The bottom of the ladder, on a path that works

  2. The rung most people skip: is it even there

  3. Names, and two failures that are not the same

  4. Ping is not a test

  5. The port, and the thing behind the port

  6. Putting both machines back

Official sources