Nothing in home hosting is more typical: the site ran beautifully for weeks, the router restarted or the ISP rotated addresses overnight, and now the domain leads nowhere. The fix is rarely where people look first. Work these layers in order and you will find the broken one in minutes, not evenings.
A home-hosted website is a chain: domain, DNS record, router, local address, server software. An IP change stresses the first three links at once, which is why the failure feels mysterious. It is not. Each link can be tested in isolation with one command or one glance, and this guide takes them in the order they actually break - starting with the layer an address change hits directly.
Ask DNS what it is telling the world, then ask the world what your address is:
nslookup yourname.example-dns.com (an example hostname) - this is the address visitors are being sent to.A stale record means the thing that rewrites it has stopped. Check its three usual hiding places:
Fix what you find, restart the router to force a fresh update cycle, then repeat step one until the record shows the new address.
Public side correct, site still dead: now the break is indoors. The router forwards ports to a local address - and if the PC's local address drifted (no DHCP reservation was set), the router is faithfully forwarding visitors to an empty chair, or worse, to a different device. Run ipconfig on the hosting PC, compare with the forwarding rule's target, and correct whichever one is wrong. Then set the reservation properly this time, so this exact failure can never repeat.
http://localhost; if it fails there too, start the server and set it to start with Windows or run as a service, permanently.Final test, and it is the only one that counts: a phone on mobile data with Wi-Fi off. Home routers frequently cannot loop their own public address back inside (a missing NAT loopback feature), so testing from the sofa can report a healthy site as dead - or, after DNS changes, show you a cached ghost of it. The visitor's view is the verdict. If the phone loads the site, you are done; write down which layer broke, because the note will save you the whole search next time.
Watch the video to rebuild any layer you found broken - each part of the setup is shown live, step by step.
▶️ Watch Full VideoThe local layers: the PC local address drifted away from the forwarding rule, the web server did not auto-start after a reboot, or a firewall change is blocking the ports. Also give DNS caches a few minutes after the record corrects itself - toggling airplane mode on your phone forces it to ask again instead of remembering the old answer.
Four habits remove nearly every repeat: put the updater in the router where it sees changes first, reserve the PC local address so forwarding cannot drift, set the web server to start automatically, and once a month open your own site from mobile data as a health check. Maintain those and a dynamic address behaves, in practice, like a static one.
Not really - the pool belongs to the ISP and the rotation is their design; restarting less may stretch an address's life on some networks but guarantees nothing. A paid static IP ends the topic completely. Until then, the updater discipline in this guide is what makes the change a non-event.