Network Troubleshooting Flowchart
The method (follow the arrows)
problem reported
→ 1. GATHER who? since when? error text? scope: one user or many?
→ 2. ISOLATE what still works vs what does not (the boundary is the clue)
→ 3. WHAT CHANGED? update, new cable, new config, storm last night?
→ 4. PROBABLE CAUSE — test the cheapest one first
└─ out of your authority or depth? → ESCALATE (hand over notes)
→ 5. FIX ONE THING (plan the rollback before you touch anything)
→ 6. VERIFY — did the original complaint actually go away?
└─ no → undo, back to 4 with the next cause
→ 7. DOCUMENT symptom, cause, fix — tomorrow's you will search for it
The bottom-up OSI ladder
Start at the floor — the cheap fixes live there. Each rung: run the check; if it passes, climb.
- Physical — link lights on? cable seated, undamaged, right port, PoE budget OK? No light → this rung is your problem.
- Link — interface up, speed/duplex sane? Win:
Get-NetAdapter· Linux:ip link. - Ping yourself — Win:
ipconfig /all· Linux:ip addr. Valid address? Thenping 127.0.0.1and your own IP. 169.254.x.x → stop, DHCP never answered. - Ping the gateway —
ping <gateway>(find it inipconfig/ip route). Fails → local network or gateway problem, not "the Internet". - Ping a remote IP —
ping 1.1.1.1; if it dies mid-path,tracert/traceroute(ormtr) shows where. - Test DNS —
nslookup example.com/dig example.com. IP ping works but names fail → it was DNS. - Test the port/service — Win:
Test-NetConnection host -Port 443(with-Portit tests TCP only and does not ping) · Linux:nc -zv host 443. - Application — the network is clean; look at config, credentials, logs.
Symptom → most probable cause
| Symptom | Look at |
|---|---|
| Address is 169.254.x.x (APIPA) | DHCP unreachable: server down, wrong VLAN, dead cable |
| Ping own IP OK, gateway fails | wrong gateway/mask setting, VLAN membership, or the gateway itself |
| Ping 1.1.1.1 OK, names fail | DNS — wrong server configured, or the server is down |
| Intermittent Wi-Fi | interference, weak coverage, channel congestion, roaming between APs |
| One whole VLAN dark | trunk not carrying that VLAN, its gateway (SVI) down, no DHCP scope |
| "Duplicate IP" warnings | a static address inside the DHCP scope, or two statics colliding |
| Same-site hosts: some reachable, some not | a wrong subnet mask splitting one network in two |
| Port dead after a device swap | port security shut it down (unknown MAC violation) |
| Whole segment crawling, switch LEDs solid | broadcast storm — a loop without STP, or a jabbering NIC |
| "Connected, no Internet", browser redirected | captive portal waiting to be accepted |
Escalate when: the fix is outside your authority (ISP circuit, another team's server) · you are out of depth or the clock/SLA is burning · the next test risks damage. Hand over everything you gathered, tried and ruled out — that work is never wasted. Escalation is a skill, not a failure.
Network Essentials · Turning Point Academy — backbone: Al-Doori, Network Essentials, Ch. 12 & 2.