Ethernet Not Working? Step-by-Step Troubleshooting Guide
If your Ethernet is not working, you've connected the cable, the router appears to be running, and yet your computer says Ethernet is unplugged or refuses to go online. Before you replace the router or call your Internet provider, stop and identify where the connection fails. A dead cable, disabled adapter, damaged wall jack, incorrect switch port, missing driver, and upstream outage can produce similar symptoms but require very different fixes.
The quickest diagnosis follows the path from the computer outward: physical link, network adapter, local address, gateway, and then the Internet. That order prevents wasted resets and helps Edmonton home users, remote workers, and small offices decide whether they need a replacement cable, a network change, or on-site technical help.
Table of Contents
- Why Your Ethernet Connection Might Be Down
- Checking the Physical Connection and Cabling
- Resetting Network Adapters and Drivers
- Interpreting IP Addresses and Gateway Results
- Avoiding Unsafe Workarounds and Security Risks
- When to Call in On-Site Professional Help
Why Your Ethernet Connection Might Be Down
A common Edmonton support call starts with a single desktop that worked yesterday and now displays “network cable unplugged.” Wi-Fi still works on a phone, the router lights are on, and the user concludes that the ISP must be having an outage. In practice, that conclusion is often premature.
The CRTC's broadband measurements provide useful context. In its October 2019 fixed wired testing, approximately 5,000 Canadian volunteers measured connection speed, latency, packet loss, and web-page loading across regions and technologies. The vast majority of participating Internet service providers met or exceeded their advertised maximum download and upload speeds, while average download performance exceeded 100% of advertised speed across all measured regions and technologies. Peak-hour congestion generally reduced download performance by only about 1–2 percentage points, a change the regulator considered unlikely to be noticeable to users. See the CRTC broadband measurement results for the methodology and findings.
That doesn't prove your provider is faultless. It does establish a sensible starting point: when one Edmonton computer reports that Ethernet is unavailable, inspect the local connection before treating the problem as a city-wide or provider-side failure.
Practical rule: If only one computer is affected, suspect the computer, cable, wall jack, or switch port first. If every connected device loses access, investigate the router, modem, switch, provider handoff, or wider outage.
The distinction matters for businesses, too. Monitoring platforms can help administrators correlate link failures, device availability, and recurring port events. If you're comparing approaches, this overview of managed network monitoring tools is useful background, but monitoring won't repair a physically broken cable. It helps show whether the failure is isolated, repeated, or part of a larger event.
A short upstream outage can also look different from a local Ethernet failure. CRTC fixed-wireless measurements found that outages could occur frequently but were generally brief, with major outages lasting more than three minutes representing 2% of recorded outages. That fixed-wireless result isn't a measurement of your computer's Ethernet port, but it reinforces the diagnostic point: a momentary Internet interruption isn't the same as a persistent “cable unplugged” message. The CRTC's fixed-wireless reliability results provide that broader Canadian context.
Checking the Physical Connection and Cabling
Start with the part you can see and touch. Software troubleshooting can't restore a link when the cable's plug isn't making contact, the wall jack has failed, or the switch port is damaged.
Follow the cable from the computer to the wall or network equipment. Look for tight bends, crushed sections, pulled connectors, damaged jacketing, and plugs whose retaining clips have snapped. Unplug and firmly reconnect both ends. A connector can appear inserted while the contacts remain slightly short of the jack's contact points.
Use link lights as your first clue
Check the Ethernet port on the computer and the corresponding router, switch, or wall-jack equipment. Depending on the hardware, a steady light indicates a negotiated link while a blinking light usually indicates traffic. Light colours vary, so don't interpret a particular colour as a universal fault code. The important observation is whether either end detects a physical connection.
If both ends remain dark, test in this order:
- Replace the cable with a known-good, factory-made cable. A cable can pass a basic continuity check and still have incorrect pair wiring, poor termination, internal damage, or unsuitable conductors.
- Move the computer to another router or switch port. This separates a bad cable from a failed port.
- Bypass a wall jack or patch panel temporarily. Connect the computer directly to the router or switch with the known-good cable.
- Test the original cable with another device. Keep the result written down instead of changing several items at once.
Don't spend time changing adapter settings while both link indicators are dark. The IEEE Ethernet troubleshooting guidance prioritizes cable condition, link lights, addressing, hop-by-hop testing, and switch verification for a reason. Physical-layer evidence narrows the fault faster than repeated reboots.
If a cable was terminated by hand, check that the conductor pairs follow the correct wiring pattern at both ends and that the plug matches the cable type. A multimeter continuity test alone may miss a split pair or incorrect pin mapping. For an infrequently used home setup, buying a reliable pre-terminated cable is usually a better trade-off than repeatedly crimping questionable plugs.
Resetting Network Adapters and Drivers
A lit port doesn't guarantee that the operating system is managing the adapter correctly. Updates, sleep cycles, dock changes, and sudden power loss can leave a driver in an unusable state even when the cable and switch port are healthy.
Windows adapter recovery
On Windows, open Device Manager, expand Network adapters, and identify the wired Ethernet device. Right-click it, choose Disable device, wait briefly, and choose Enable device. This reloads the device state without changing your network design.
If that doesn't help, open the adapter's Properties and check the device status. A warning icon points toward a driver or hardware problem. Use Windows Update or the computer manufacturer's support page to install the correct driver for the model. Avoid random driver-download sites, especially on a business computer.
Windows also includes a network troubleshooter. Run it after confirming the cable and port, not instead of checking them. The troubleshooter can repair common adapter and TCP/IP settings, but it can't correct a damaged wall jack or an incorrect business VLAN.
macOS network service reset
On a Mac, open System Settings, select Network, and choose the Ethernet service. Confirm that the service is enabled and that the correct physical interface is selected, especially if the Mac uses a USB-C dock or Thunderbolt adapter.
Remove and re-add the Ethernet service if its status appears stuck. This forces macOS to rebuild the service association while preserving the rest of the computer. If the problem began after sleep or a system change, shut down fully, disconnect the dock or adapter, restart, and reconnect it directly where possible.
macOS can also retain stale network state in NVRAM. Resetting NVRAM is a maintenance step, not a first response, and the exact startup procedure depends on the Mac model. Record custom display, startup, or security settings before using it, because a reset can affect more than Ethernet.
Linux driver and service checks
Linux users should first identify the wired interface and inspect its state with the distribution's network settings or NetworkManager. A disconnected interface can often be re-enabled from the desktop network menu.
For a system using NetworkManager, restarting the service can refresh the connection:
sudo systemctl restart NetworkManager
If the adapter driver itself is stuck, unload and reload the relevant kernel module only when you know its name and have another way to recover access. A safer first move is a full shutdown, removal of USB network hardware, and restart. On a business Linux workstation, document the current driver and interface state before changing modules.
The internal network adapter reset guide can help you follow the appropriate platform path. Don't combine a driver change, operating-system reset, and router reboot in one attempt. Make one change, test the link, and record the result.
Interpreting IP Addresses and Gateway Results
Once the physical link and adapter are active, ask a more precise question than “Is the Internet working?” Ask which device your computer can reach.
A computer that receives a self-assigned APIPA address beginning with 169.254 generally hasn't reached a DHCP server. That points toward the local adapter, cable path, switch port, VLAN, or DHCP service. It doesn't automatically prove that the ISP is down. Check whether a known-good computer on the same port receives a normal local address before changing router settings.
A valid local address with no gateway response suggests a different fault. The device may have a bad subnet mask, incorrect manually assigned settings, a switch configuration problem, a firewall rule, or a wrong VLAN. In a small office, a port-security change or a recent move to a different wall jack can create this symptom without any hardware failure.
Test one hop at a time
Use the operating system's network diagnostics or ping utility in a controlled sequence:
- Test the local interface. This checks whether the computer's network stack can communicate with itself.
- Test the directly connected switch or router. Failure here keeps the investigation on the local link, adapter, port, VLAN, or addressing.
- Test the default gateway. If the gateway responds, the computer has at least reached the local routing boundary.
- Test an external IP address. Success here with website-name failures points toward DNS rather than Ethernet.
- Compare with a known-good device. Use the same cable, port, and router wherever possible.
The IEEE Ethernet networking report recommends this cable, link-light, addressing, and hop-by-hop progression, along with checking switch VLAN and speed or duplex settings. Its value is simple: each failed hop removes entire categories of possible causes. The IEEE troubleshooting report describes the sequence in more detail.
The IP address assignment guide is relevant when a device has an unexpected local address or fails to obtain one. Don't force a manual address just to make the warning disappear. That can mask a DHCP or VLAN problem and create a second conflict on the network.
For managed networks, record the adapter address, assigned VLAN, negotiated speed, duplex, switch-port errors, and test times. The Canadian Centre for Cyber Security recommends device inventories, configuration and firmware records, centralized logging, bandwidth audits, and synchronized time sources. Those records turn an intermittent complaint into evidence that can be matched against a port change, firmware event, or utilization spike.
Avoiding Unsafe Workarounds and Security Risks
A working Ethernet light only proves that the physical link negotiated. It doesn't prove that the computer is on the correct network, receiving the intended policies, or protected by the intended firewall.
That distinction matters when a user is under pressure. A personal hotspot may restore access, but it can put work traffic on an unmanaged connection. Disabling the endpoint firewall may make a test succeed while exposing the computer. Connecting directly to an ISP modem can bypass the router's firewall, address policies, monitoring, and segmentation.
Choose a temporary connection carefully
A safe workaround should preserve security while you continue diagnosing the normal path:
- Use a trusted temporary connection. A managed Wi-Fi network or approved hotspot is preferable to an unknown public network.
- Keep firewall and endpoint protection enabled. If a test appears to require disabling them, stop and involve the person responsible for the computer.
- Confirm the network role. Check whether the device is on the intended home, guest, or business VLAN.
- Avoid direct modem connections for routine work. Use one only when the network administrator understands the consequences and has a defined test plan.
- Restore the normal path after testing. Remove temporary profiles and reconnect the computer to its approved network.
The Canadian Centre for Cyber Security recommends placing a firewall at the boundary between a corporate network and the Internet, and separating wireless traffic from wired traffic with a distinct VLAN. Its guidance on basic perimeter defences supports a point that many consumer troubleshooting pages miss: wired doesn't automatically mean secure.
Authentication, segmentation, firewall filtering, switch configuration, and endpoint controls determine the security posture. A cable plugged into the wrong wall jack can provide an active connection to an unexpected network. An unmanaged switch can also be connected in a way that defeats the intended separation between guest and business traffic.
Check the logic behind a successful test
Suppose a laptop reaches a website through a hotspot. That proves the laptop can communicate through the hotspot. It doesn't prove the Ethernet adapter, office DHCP, business DNS, firewall policy, or wired VLAN works.
Suppose the Ethernet link light comes on after moving a cable. That proves the port negotiated at that moment. It doesn't verify the assigned network, gateway route, or policy enforcement. Record the local address range, gateway result, and network name before declaring the fault fixed.
If an improvised workaround is necessary for urgent work, use it briefly, keep sensitive activity within approved controls, and tell the responsible administrator what changed. Restoring connectivity is useful, but restoring it by bypassing security can leave a larger problem behind.
When to Call in On-Site Professional Help
DIY troubleshooting has a stopping point. If a known-good cable, another switch port, a direct connection, an adapter reset, and a gateway test all produce the same failure, further changes may obscure the cause. Use the results to choose the right handoff.
| What you observe | More likely fault area | Sensible next action |
|---|---|---|
| No link lights with multiple known-good cables | Computer port, wall jack, switch port, or hardware | Test another device and arrange physical inspection |
| One computer fails while others work | Adapter, driver, TCP/IP stack, firewall, or local settings | Preserve the device details and repair the affected computer |
| Several devices fail at the gateway | Router, switch, VLAN, DHCP, or ISP handoff | Check network equipment and provider status |
| Gateway works but external access fails | Routing, firewall, DNS, or upstream service | Review the router and firewall path before replacing hardware |
| Link returns and drops repeatedly | Cable termination, port errors, negotiation, power, or hardware | Capture timestamps and port counters before changing settings |
Know what on-site repair can resolve
A technician can inspect wall jacks, test cable pairs, verify switch-port negotiation, check adapter drivers, replace failed network cards, and assess router or small-office wiring. On-site diagnosis is useful when the fault changes as a cable moves, affects one room, or involves several devices on a managed network. For a full walkthrough of what a technician brings to the visit, see what on-site computer repair includes.
Hardware replacement differs from board-level repair. Replacing a failed network adapter, dock, port assembly, or other standard part is a normal repair workflow. Component-level motherboard work requires fault isolation, microscopic inspection, specialized rework such as BGA, SMD, or microsoldering, functional testing, and firmware verification. Canadian board-repair providers describe that work as a case-by-case process based on the model, board condition, parts availability, repair history, and reported failure. This explanation of board-level repair shows why board work is different from swapping a standard component.
The on-site service described here handles most major hardware repairs at the customer's location but does not provide board-level repairs. It also doesn't provide remote services, so an in-person visit fits faults involving a physical device, cable path, or local network configuration.
Decide whether you need repair or ongoing oversight
A single failed cable calls for targeted repair. Repeated failures across ports, devices, or locations may indicate poor documentation, ageing equipment, unmanaged changes, or a network that needs regular observation.
Managed service providers generally handle broader day-to-day IT work, including help-desk support, patching, backups, security services, and continuous monitoring. Ongoing network monitoring can still help without claiming full MSP coverage. It can track connected devices, availability, performance, and security events while leaving wider responsibilities outside the service boundary. The Canadian guide to managed IT services explains that distinction.
For a growing office, combine repair records with an inventory of switches, access points, adapters, wall-jack labels, firmware versions, and approved network layouts. Teams examining broader technology planning can read about digital transformation, but transformation projects will not fix a loose connector or misconfigured port.
Before booking help, record the cables and ports tested, whether link lights appeared, the adapter status, the assigned address type, gateway results, and whether another device worked on the same connection. This record lets a technician begin at the failed layer instead of repeating every basic step.
Nerds 2 You Edmonton provides on-site computer repair, wired network troubleshooting, device diagnostics, and support for home users and small businesses, while excluding remote services and board-level repairs. If Ethernet still fails after the physical and gateway checks, visit Nerds 2 You Edmonton to arrange help at your home or office.
Discover more from Nerds 2 You
Subscribe to get the latest posts sent to your email.
