I have tried to set the address of the lan interface to 192.168.0.1 to fit into an established network. The setting is saved and the interface issues dhcp addresses on the 192.168.0.0 network but I cannot login to the device by ssh, or web nor can I ping the interface. This is on a Ubiquiti UAP AC PRO Gen2. I have set other devices lan addresses to the 192.168.0.0 network with openwrt but this one will not allow access after the change even though it will issue dhcp leases. Openwrt version is openwrt-23.05.0-ath79-generic-ubnt_unifiac-pro-squashfs-sysupgrade.bin. Any help would be greatly appreciated. I really don't want to change the address of every device on the network.
If you are using a /24 network, the .0 address is invalid. Is that the address you actually set?
If so, you will need to use failsafe mode to edit the file to fix the invalid address or reset to defaults and start over.
The above post clearly states that I am assigning the 192.168.0.1 address from the 192.168.0.0 'network' to the lan interface. I have also tried 192.168.0.250 and 192.168.0.254 with the same failure. if I assign 192.168.2.1 everything works fine. As stated I have assigned 192.168.0.1 to other devices in the past running OpenWrt (various linksys routers) with no issue so I am wondering if there is a possible issue with the Ubiquiti AC PRO or it's OpenWrt binary.
Hello! Would you mind posting the contents of the following files from your system? (After setting your config back to the 192.168.0.0/24 net) Feel free to redact any sensitive addresses.
/etc/config/network
/etc/config/dhcp
/etc/config/firewall
And also the output of ifconfig, please?
1 Like
That's a problem. I cannot access the AP after I set the network. No ssh no webui. Interestingly, if I set the address to 192.168.2.4 everything works perfectly but if I do the exact same setup using the 192.168.0.1 address it will serve DHCP leases but cannot be logged in to.
Sorry... I was reading quickly and got slightly confused about the situation.
How have you been making the change? (LuCI, UCI config cli, direct editing of the config files?) Have you been immediately renewing the DHCP lease after making the change? What does the DHCP lease look like (what address, subnet, router, and dns information is provided)?
Also, are you positive that there are no IP address collisions? What about other DHCP servers?
Also, how are you regaining access to the device after you can't access it at the 192.168.0.1 address?
Very unlikely.
Let's see the config files (from a working state on 192.168.2.4)?
Please connect to your OpenWrt device using ssh and copy the output of the following commands and post it here using the "Preformatted text </> " button:
Remember to redact passwords, MAC addresses and any public IP addresses you may have:
ubus call system board
cat /etc/config/network
cat /etc/config/wireless
cat /etc/config/dhcp
cat /etc/config/firewall
2 Likes
Please help me out and reconcile those, seemingly incongruent, statements
Agreed, it is in the 16 bit block. So 192.168.0.0/16
The OP never actually told us the subnet mask, but they did confirm that they did not set the .0 address. That said, since it is the beginning of the block, there is no subnet size where that .0 would be valid.... but the OP has said that the actual address used was 192.168.0.1.
This is where I failed ip addressing: I cannot do binary octets. All I know is those last two 0s are 00000000.
Well, anything that's a /12 or bigger, as I recall, it could be valid (because it's not the top of the subnet), but not used in practice.
The fact that assigning 192.168.2.4/24 (?) works is actually interesting. Seeking the configs will help solve this mystery.
We need to better understand this "established network" as well.
The OP did make some conflicting statements, or perhaps we need clarification. It makes me wanna verify that the OP: doesn't have another device issuing DHCP leases (and did the OP make sure everything got a new lease), what network/interface the AP connects with, network and firewall configs, etc.
The Op notes:
This conflicts.
1 Like
The bold part gives it away I think 
1 Like
Yeah, I had to read that about 5 times and decided he has one device at present but has had other OpenWrt builds work in this present configuration.
I am interested in seeing how their devices' configurations look after being configured by the router.
I have a number of HA devices which need static leases so that their location is reliable. The static leases are handled by the dhcp server running in openwrt on a linksys wrt1900ac which functions just fine on the 192.168.0.0/24 network. This is a ubiquiti UAPAC PRO that is simply supposed to be an access point. I added the fact that it could issue leases as a data point that some part of the config is working. I have tried to config with UCI cli, direct editing of /etc/config/network and with LUCI with the same results in each. When I try to change the lan address in LUCI I need to log back in within 90 sec or it rolls back, I change the lan address then restart the laptop interface to pull a new address and try to log in. That is where the problem is, while the dhcp server will issue the laptop a lease at 192.168.0.142 I cannot login to the AP. This is all being done in complete isolation, laptop, AP and poe injector before placing the AP in service in my dad's network to solve some dead spots in the house. For reference I have over 100 static leases in my network for wifi lights, plugs and appliances that NEED to be at specific addresses and would really rather not edit all those leases to a new subnet. I appreciate the help but I am not able to login to the AP once I change the lan address by ssh or web(LUCI) interface. As I stated, if I set the lan address to 192.168.2.4 or 192.168.2.1 (I have tried both of these) all works fine, if I ONLY change the lan address to 192.168.0.1 or 192.168.0 250 (I have tried both) then I cannot login. The address change is the ONLY config change so I am lost as to why I cannot login on the 0 subnet but ok on the 2 subnet AND the WRT1900AC does work and can be logged into on the 0 subnet. Also there is not a WAN interface, I ONLY have a lan interface and wifi bridge to the lan. I am thinking that the 0 subnet is maybe reserved or possibly openwrt thinks it is the WAN address and is blocking access for security? Just my best guesses.
Just to be clear, you do understand that you cannot have 2 DHCP servers running on the same network, correct?
If isolated then, please explain if these devices work when you configure them.
To be clear, are you saying that you know you setup some kinda IP mismatch - because you don't want to edit the static lease config in /etc/config/dhcp?
Also, explain how the issue arises if you tested only in isolation?
This is why we need to see the config. We can't guess your error or misconfiguration.
And more importantly, you cannot login...from where, what device, etc. (e.g. what SRC IP)? 
Can you clarify which network number is the "established one", which one has a list of leases, etc.?
So you have a DHCP server on the 192.168.0.0/24 network already?
What are you doing for the change of address? Are you setting a different static IP or are you changing the protocol to DHCP client? It sounds like you're doing a static IP change -- is that correct?
How are you connected to the AP? Based on this, it sounds like you're connecting via ethernet to the device, correct (computer > PoE injector > AP main ethernet port)
Is your laptop's wifi disabled? If not, this will likely cause a problem since you already have a 192.168.0.0/24 subnet. There should be literally no other network connections, wired or wireless on your laptop.
I believe that you are running into a conflict because you have another network connection on your laptop -- likely wifi.
No, this is incorrect. There is no reservation or other special treatment of the .0 subnet for OpenWrt and/or your UAP-AC-PRO.
1 Like
You have two DHCP servers and/or two devices assigned to 192.168.0.1
You can run arp -n to see the MAC associated with 192.168.0.1
Turn the router on and ping 192.168.0.1 several times then turn it off and ping and check MAC again.
If you see the MAC change then those (two MACs) are the two nodes both assigned to 192.168.0.1
Sometimes a node will not respond to the pings but they have to respond to the ARP so even if the ping fails still check for a MAC change.
If you have a Wireshark setup then you can issue an ARP and look for multiple responses.
1 Like
Yes, I am trying to set the lan address to static. The laptop is not connected to anything but the poe injector. The other openwrt device that will be serving dhcp is currently 250 miles away so only one device serving leases and only 1 device on 192.168.0.1. There is no IP mismatch, I have an established network on the 192.168.0.0/24 network and want to add an access point to cover a dead area in the house hence I need to place the AP in the existing 192.168.0.0/24 address space. The ONLY reason I have the dhcp server running on the AP right now is to give the laptop a lease as I got tired of having to race to change the interface manually in the 90 seconds, much easier to disconnect, reconnect, pull new lease. The AP will not be serving dhcp once placed into the network but since I cannot login to the AP after the lan address change I cannot get to it to check the ignore interface box. I am unable to login to the AP after changing the lan address, the AP dhcp gives my laptop the address 192.168.0.142, the AP lan address is set to static, 192.168.0.1 and I am trying to login over the wired interface thru the poe injector. Again ALL of this works if I set the address to 192.168.2.4 but not if I set to 192.168.0.1. I will edit the config files and post them but it will have to be before the reboot as I cannot get to them after the reboot.
So how are you getting back into the device after you loose contact with it? Are you resetting it?
Please post your configs as requested earlier.
1 Like
Yes hard reset. Well, I found the problem. I don't use the laptop much and for some reason the last time I used it I had configured a sub interface statically at 192.168.0.10 in /etc/network/interfaces. This did not show up in network manager so I did not realize that the laptop had two routes to the AP after the lan address change to 192.168.0.1. Once I deleted the sub interfaces all is right with the world. It's been so long since I used the laptop I don't even remember why I set those sub intgerfaces. I thank you for your patience with me while I slink away with my tail between my legs.