Router service recovery through a Tailscale jump

Today a router-side service that helps me reach home went down. The awkward part was that this service is also part of the path I normally use to get back home, so direct access to the router was unavailable from outside.

The useful foothold was a small home server that was still online as a Tailscale node. It could still see the router on the home LAN. That made the recovery path simple: connect to the independent Tailscale node, then jump from there to the router's local management address.

ssh -J user@tailscale-node admin@router-lan-address

Once on the router, the fix was not a rebuild or a config migration. It was just the boring manual action that mattered: restart the service and confirm it came back.

/etc/init.d/network-service restart
/etc/init.d/network-service status

The important lesson is not the exact IP pair. It is the shape of the access path: keep at least one independent Tailscale node inside the home network that is not the router itself. If the router's proxy or exit path fails, that node can still act as a jump host for emergency maintenance.

This is also a reminder that "remote access works" should not mean "the router is the only way in." The router is infrastructure under repair; it should not be the only tool available to repair itself.