Get One Month of Starlink FreeSPECIAL OFFER: GET A FREE MONTH OF STARLINK >
Connected network ports on a home router

Test Starlink Dual-WAN Failover and Failback Without False Alarms

Tips & Tricks

Back to Tips & Tricks

A dual-WAN setup is only useful if it changes over cleanly when Starlink fails and returns to the preferred path when Starlink recovers. A page loading after you unplug one cable is not enough: the browser may be showing a cached page, the router may still think the wrong link is healthy, or a device may have switched to cellular.

Use this guide to test failover and failback in a controlled way. It applies to a Starlink connection paired with fibre, cable, 4G/5G, or a second Starlink connection. Menu names differ by router, so use the equivalent controls in your own router.

Before you start

Choose a quiet test window. Avoid testing during an important call, a remote support session, a payment, or a firmware update.

Record these details before changing anything:

  • Which WAN is preferred during normal operation.
  • Which WAN is the backup.
  • The router's health-check target and failure threshold.
  • Whether the router is in router mode or access-point mode.
  • The device you will use for testing.
  • The public-facing address or connection name shown by the router, if available.

Keep one device connected to the same local network throughout the test. A second device on mobile data is useful for checking whether the problem is local to your home network.

Open a simple live service, such as a page that shows the current time, and start a continuous ping to a stable host if you are comfortable reading the results. A successful speed test is not required for this test; continuity, route selection and recovery are the important signals.

Confirm the normal path

  1. Check that the router reports Starlink as the active or preferred WAN.
  2. Load two different websites and start a small, non-critical download.
  3. Note the time and the router's WAN status.
  4. If the router shows a public address or connection label, record it. Do not treat an address change alone as proof of failure or success.
  5. Confirm that local devices can still reach the services they need, such as a printer, NAS or home-automation controller.

If the normal path is not healthy, stop here and fix the baseline first. A failover test cannot tell you whether a problem was caused by the primary link or was already present.

Test failover

Use the least disruptive action your router provides first. Disabling the Starlink WAN in the router is usually easier to reverse than removing power from the terminal.

  1. Start a timer and note the current time.
  2. Disable the Starlink WAN, disconnect its Ethernet path, or use the router's built-in test control. Change one thing only.
  3. Watch the router's status page. It should mark the Starlink path unhealthy and select the backup WAN after its configured checks fail.
  4. Refresh the two websites and start a new small download. A fresh request is more useful than an already-open page.
  5. Check the continuous ping, if you started one. A short interruption is expected; repeated long gaps or a permanent failure needs investigation.
  6. Confirm that a device still reaches the local services you recorded earlier.
  7. Leave the backup active for several minutes and verify that the router does not oscillate between WANs.

The result is a pass only when the router changes to the backup path and new traffic uses it. Do not call it a pass because the Wi-Fi icon stayed visible: Wi-Fi can remain connected while the internet path is broken.

Test failback

Failback is the return to the preferred path after Starlink is healthy again. It deserves its own test because some routers stay on the backup indefinitely, while others switch back too quickly and create a flapping connection.

  1. Reconnect the Starlink path or re-enable its WAN.
  2. Wait for the router's health checks to report the link healthy. Use the router's configured recovery or hold-down interval if it has one.
  3. Confirm that the router returns to the intended preferred WAN. If you want the backup to remain primary until a scheduled change, confirm that this is what your policy says instead.
  4. Start a new web request and a new small download. Check that traffic behaves normally.
  5. Watch the WAN log for several minutes. There should not be a repeating cycle of Starlink healthy, Starlink failed, and backup selected.
  6. If the router displays both link states, confirm that the backup is now standby or secondary rather than still carrying all traffic.

A successful failback does not require every existing TCP or video-call session to survive. Changing WANs can change the route and the source address, so some sessions will reconnect. The useful question is whether new connections follow the intended policy and whether the network settles.

Repeat with a different failure

A cable disconnect, a Starlink service interruption and a powered-off terminal are different failure modes. After the first test passes, repeat a lower-risk version with one of these conditions if your equipment allows it.

  • WAN-path failure: disconnect only the cable between Starlink and the router.
  • Terminal failure: use the normal shutdown or power control for the Starlink equipment.
  • Router detection failure: leave the cable connected but confirm whether the health-check target is reachable through the wrong path.

Do not pull outdoor cables or open equipment to simulate a fault. Stop if a connector, power supply or cable looks damaged.

What to record

Keep a short test note with:

  • Date and time.
  • Primary and backup WAN names.
  • The action used to simulate failure.
  • Time until the router selected the backup.
  • Whether new web requests worked.
  • Whether local services stayed reachable.
  • Time until the preferred path returned.
  • Any log message, alert or repeated WAN change.

This turns a vague claim such as “the backup works” into a repeatable result. It also gives you useful evidence when tuning health-check intervals or asking the router manufacturer for help.

Common false alarms

A cached page still loads. Open a new private window or request a page you have not just viewed.

The Wi-Fi signal is strong but the internet is unavailable. Wi-Fi only proves the device can reach the local router. Check the router's WAN state.

A speed test passes once. A speed test can use a different server or route. Test new connections, continuity and the router's selected WAN.

The router never fails over. Check that the health check is testing an internet destination and that the backup WAN is actually configured as eligible. A monitor that can be reached through the local router may report success even when Starlink is unusable.

The router flaps between paths. The health threshold may be too sensitive for a variable link. Increase the required failed checks or recovery hold time carefully, then repeat the test.

Failback never happens. The router may be configured for manual return, a long recovery interval or a backup-preferred policy. Verify the intended policy before changing it.

Practical pass criteria

Your setup is ready for real use when:

  • Starlink is preferred during normal operation.
  • One controlled Starlink failure selects the backup without manual router reconfiguration.
  • New web requests work over the backup.
  • Local network services remain available as intended.
  • Starlink's recovery returns the system to the intended policy.
  • The router remains stable for several minutes after recovery.
  • You have a short record of the test and know how to reverse the simulated failure.

Restore all cables and router settings after testing. Never leave a WAN disabled or a power connection loose simply because the backup test passed.