Symptom	Layer	Check
Can't ping 192.168.100.1 at all	LAN / router config	New router likely defaulted to a different subnet — revisit ask 1
Local test passes,	Port forwarding	Re-check the two NAep 2)
DNS still shows the old IP after 5+ min	DNS / cron	~/duckdns/duck.sh, then crontab -l and catd>
Resolves and connects, but shows 502	Docker containersdocker ps -a — restart whichever isn't healthy
Everything works bu>	Switch port	Confirm router and t 9 or 10, not 1–8

Docs

IP Change Runbook

Docs Guides

When the Home IP Changes

The recovery sequence for Nextcloud and the WordPress sites whenever the ISP-issued public IP changes — written for tomorrow’s router replacement, but good for any future IP change too.

Covers: ahaddad.duckdns.org & ahaddad-wp.duckdns.orgServer: 192.168.100.13Last verified: 2026-08-21

ubuntu-serv>

When the Home IP Changes

The recovery seqordPress sites whenever theISP-issued public IP changes — written for the router replacement, but good for any future IP change too.

Covers: ahaddad.duckwp.duckdns.org Server: 192.168.100.13 Last verified: 2026-

Solid line: the path a request actually takes. Dashed line: the separate control loop that keeps the hostname pointethis now runs on its own via cron,every 5 minutes.

Before the technician leaves

Ask them to keep the LAN on 192.168.100.0/24, gateway 192.168.100.1.<

Every static address in the house — the server, the switch’s own admin IP — assumes exactly this ranto 192.168.1.0/24 out of the box; if the new one does, it breaks the whole LAN at once, not just the sites.

2

Plug the new router i>port 9 or 10 — never1–8.

Ports 1&ndasware-limited to 100Mb, permanently,no setting fixes it. Landing the router there silently caps your whole connection again, exactly like before.

Recovery sequence, in order

1

Confirm the server still has a network

Checks the pfails, nothing below it can work— and the fix is the LAN/router setup, not DNS or Docker.

R
ip -br addr show usb-eth-hub ping -c 3 192.168.100.1 ping -c 3 8.8.8.8 # the internet

&#10003ucceed. If the router ping fails,the LAN subnet likely doesn’t match — revisit ask 1.

2

Re-create port forwardi
The old router’s forwarding rules live inside that specific device. A full replacement means a blank routash; DNS can be perfect and theserver perfectly healthy, and the outside world still can’t get in without this.

Aouter’s admin page

External 80/tcp → 192.168.100.13:80 External 443/tcp → 192.168.1


      3
      
        Let the DNS auto-update happen — or trigger it yourself
        Both domains public IP before anyone, includingyou, can reach the sites by name. A cron job runs duck.sh every 5 minutes
and self-corrects.
        Just wait up to 5 minutes. To check or force it sooner
        curl    # today'sactual public IP
getent hosts ahaddad.duckdns.org   S currently says
~/duckdns/duck.sh                    # force an update right
now
        ✓ Expect: the two values match. If they still don't after
5+ minutes, see the reference table
      
    

    
      4
      
        Verify the sites are actually reachable
        Each test be; DNS, port-forward, or the appitself — so a failure tells you exactly where to look.
        Lely
        curl -sk -o /dev/null -w "HTTP %{http_code}\n" \
  --resolve ahaddad.duckdns.org:443
  https://ahaddad.duckdns.org/
        ✓"mono">HTTP 302
        External test
        Open  and https://ahaddad-wp.duckdns.org/ in a browser, ideally on a phone off Wi-Fi (real
outside path).
        If anything looks wrong, check container health
        dockames}}\t{{.Status}}"
        ✓ Expect: every row reads Up /
(healthy)
      
    

    
      5
      
        Nextcloud's trusted-domains list — no action needed
        Nextcloud one it's accessed by, ahaddad.duckdns.org, never the IP behind it. That hostname doesn't change, only
what it resolves to. This step is hi> to go looking for something tofix inside Nextcloud.

      

    

  


  
    If something's still broken
    
      
        
          Symptom Layer Check
        
        
          
            Can't ping 192.168.100.1 at all
            LAN / router config
            New router likely defaulted to a different subnet — revisit ask 1
          
          
            Local test passes,
            Port forwarding
            Re-check the two NAep 2)
          
          
            DNS still shows the old IP after 5+ min
            DNS / cron
            ~/duckdns/duck.sh, then crontab
-l and catd>
          
          
            Resolves and connects, but shows 502
            Docker containersdocker ps -a — restart whichever isn't healthy
          
          
            Everything works bu>
            Switch port
            Confirm router and t 9 or 10, not 1–8
          
        
      
    
  

  ubuntu-server · home network runbook · written 2026-08-21