Hello, when I configure access to the modem/ont at ip 192.168.2.1, with the ip being fixed at 192.168.2.2 and the network mask at 255.255.255.0, the ddns starts pulling this ip from the WAN, whereas it should pull the external ip. As a result, I can’t access my modem/ont without damaging the ddns, could someone help me if this is a bug? because I do the same configuration in pure openwrt and I don’t have this problem with ddns
It’s not a bug, it’s by design how the DDNS service works.
The DDNS does not make a query to an external service, it will just use the WAN1 address.
Pure OpenWRT does not use the GL DDNS service, that’s why it is working there.
So, I can’t access my modem using glddns, is there no other way around this?
In a normal setup, you wouldn’t be able to access your modem anyway, so I would assume yes, there is no easy way around this. Why do you need to access your modem that often?
I think I control my internal network, always keeping track of my fiber signal if it is within acceptable parameters.
And just trusting that it’s OK isn’t an idea? 
You could try to use another DDNS service by installing the packages from OpenWrt
yes, and a way to use it, OK thanks.
You still can bind a wan address of your choice with GLDDNS, thus bypassing inquiring an IP from the wan interface!
Check the steps I listed in this thread;
1 Like
But is this procedure for fixed or dynamic IP? Because the point of using DDNS for me is because my ISP uses dynamic IP.
Edit:Disregard, I’m reading the whole topic, and that’s exactly the mistake I have! I am going to try!
It didn’t work, I played the commands as instructed, but the error still appears in glddns pulling the fixed IP of my operator’s modem.
Did I do something wrong?
Even with the error of pulling the fixed IP that I put, the GL DDNS updated it to the new real IP.
You first enable the ddns from the web gui. Then you follow the commands I listed to override the bound ip in GL DDNS server.
But it will only work once, so you need to create a cronjob for that, am I right?
No, not needed as the binding is performed once when the switch is on.
But if your ISP uses a dynamic IP - how will this change be populated to the GLDDNS?
It does not matter - binding from the shell will stay forever unless the switch is off and on again.
You don’t get my point, I guess.
For example in Germany:
ISP provides you with an router + dynamic IPv4 address for this router. (Which might renew every 24h)
Your Flint is behind this router, so the Flint get’s the WAN address 192.168.178.2 for example.
Now I use your script to push my real IPv4 address to DDNS. But the IPv4 there will be only valid as long as my ISP will not provide me with a new one. So in worst case 24h.
In this case, yes. 
If his upstream gets an updated ip then he needs to re execute the command or create some job to detect it and bind it again.
1 Like
huuum so basically this process is for fixed IP, dynamic it doesn’t make sense to keep placing the commands every time.