Ok, I couldn’t find a DUID in the network file, only dhcp_default_duid. So I backedup the file and removed it and rebooted. Did not fix it.
Restored the original network file and rebooted. Flashed 24.10.5 again… No internet. Emotional reaction But rebooted twice: internet works again
I’m tempted to flash 25.12 and reboot it some times. But tomorrow i have to work from home, so i cannot afford a non working router. I will try in the weekend.
This will be my second option to try. I guess it is a good idea anyways to keep a usb stick with a copy of a working config as well. To be able to try things first.
JonSnow, Chrisss and myself have the same type of issue. After upgrade to 25.12, No internet, wan IPv4 adress is not assigned anymore. In the main thread this advise was given and may potentially still help.
v24.10 does not send this option as part of it's wan DHCP request.
Asking 'my friend' claude_ai how to stop v25.12 to sending option 61 as part of it's wan DHCP request, it suggests the following:
Edit `/lib/netifd/proto/dhcp.sh` and remove lines 68–69: (https://forum.openwrt.org/t/problem-with-dhcp-client-and-forced-clientid-in-v25-12/247164)
# REMOVE these two lines (25.12 behavior):
[ -z "$clientid" ] && clientid="$(proto_dhcp_get_default_clientid "$iface")"
[ -n "$clientid" ] && clientid="-x 0x3d:${clientid//:/}"
# REPLACE with this single line (restores 24.10 behavior):
[ -n "$clientid" ] && clientid="-x 0x3d:${clientid//:/}" || clientid="-C"
This restores the old behaviour: if no `clientid` is set in UCI, udhcpc is called with `-C` (don't send option 61 at all).
After saving and rebooting, I'm not seeing option 61 sent as part of the DHCP request packet. RESULT!
Finally, I've not tested this as I can't afford to interrupt internet at the moment, but @mmg reports that this works with his ISP, so I have high hopes!
If anyone else tries, please let us all know how you get on.
I’m smiling while typing this. Adjusted the dhcp.sh and did a reboot. A restart of the interface was not enough. Now the wan interface gets an ip from the isp. And this is typed over the connection provided by the 25.12.
@hauke is the author of the release notes. I am happy to edit the release notes (on the forum), but only with permission. (I have concerns that the notes would be inconsistent between the forum and the other places that the release notes are posted).
Other options (not mutually exclusive):
We could create a new 'sticky' in the announcements & releases section of the forum.
We can add a wiki page that specifically discusses the mitigation for DHCP issues that seem to affect some users.
We can create a 'canned response' for use on the forums that describe the proposed solutions, point to a forum thread, and/or point to a wiki document.
Out of curiosity, is there any test that one could run to see if this issue would affect them before upgrading to 25.12? It might be helpful to include in the forum notification or sticky.
Hmm could this problem also expose different as symptom?
I still receive dhcp from the isp and get a ip, but my internet is severely malfunctioning, if I visit github or go to this forum it gets extremely slow, it halts extremely long and then loads, but if I edit a post or write a post it again hangs, if it was dns which I think is not the problem due to caching, it must be something else.
If I look in speedtest all fine, when I look to ifconfig I see no error except maybe 2 on wired interfaces but doesn't increase.
It only happens to me to recent master builds, but I wonder if the DUID change can make this weird symptom too, it feels like the tx rate is severely degraded or something does DoS, when going back to a older firmware image I made with Jenkins on 17 feb the problem was not present.
Yup, I was suspecting this too, I have changed some of the route64 to my own country and tried to change some mtu on my wireguard instances especially the cascaded ones those showed more errors in ifconfig.
But when I used no vpn it still happened, it is strange that a firmware difference alters the behaviour, I have tried to look into recent GitHub commits but i could not find a clue, I might have to do a big bisection or maybe do some more testing.
For now I have set the cascaded vpn like wifivpn/wgserver to 1280 mtu, and the normal server 1420.
I have also encountered a version of this problem and, depending on people’s ISPs, there might be a more low-tech, caveman style solution to it.
I had also noticed that after the upgrade to 25.12 my main router was receiving DHCP NAK from my ISP and couldn’t get an IP address. So I ifdown'd the interface where my wan was, waited for about 15-20 minutes and then upon running ifup the issue was gone.
I’m posting this as a potential plan A to before editing dhchp.sh . If it is out of line or obvious, I will remove this post
But rebooted twice: internet works again 
After saving and rebooting, I'm not seeing option 61 sent as part of the DHCP request packet. RESULT! 
BOOM!