Just installed HAOS and can not access

If no connection can be made with the IP address then DNS is not the first culprit here.

He most likely can ping/nslookup/tracert the IP, and homeassistant.local, if i look to the multiple previous Similar Topics ( With no or wage solutions )

And i don’t know about you, but i have no Idea how various Browsers react upon an IP-port:8123 , yes i know !, Neither 192.168.x.x:8123 nor homeassistant.local:8123 is a valid Domain, and you can’t ping/nslookup/tracert a ip:port,
However i doubt Browsers knows the Device IP-Number.it’s running on, and it have no control over the NIC, or OS for that matter, one thing i know is the message the browsers “trow” is kind of nonsense ! .

" If your computer or network is protected by a firewall or proxy, make sure that Firefox is permitted to access the web."

I can block FF, or block all internet access. ( I even have blocked my windows-machines(ip), where my “bridged” VM lives, from internet access)
( On the other hand i don’t “Force” any Devices/Apps/Programs to use External DNS’s )
And a NIC on a Windows Machine, Does what it’s configured to.
I have IPv6 disabled in Router, Windows And HA

So what if the NIC ( on a Windows machine ) prefer to use IPv6 (i.e when it’s a port which are known for various web-servers ( Microsoft “like” IPv6 ) :wink: , as well as they love to “control” your hardware

OP have IPv6 disabled in HA, But most likely not on his Windows machine (Browser host ), Or Router

If you haven’t disabled IPv6 on the Host you Browse from, you basically have no control over which IP-version is used for certain requests

i.e, Windows(with enabled IPv6) will i.e Ping and run Tracert over IPv6 ( atleast When you type in host-name(.local )/domain-name ) , so what happens when one type in an 192.168.x.x:8123 in a Browser, in windows ?
It’s not a valid IP-Number( 192.168.x.x are ), but HA don’t respond to that, Nor a valid Domain-name ( neither are homeassistant.local:8123, it’s uses mdns )

He basically should be able to just type homeassistant:8123, Hostname which Windows “also like” , And btw handles different from Domain/IP , atleast his Router should know the Hostname , and will interpret it/handle it as a Hostname, not a domain-name/ip

you might be able to telnet to it

The browser connects to port 8123 on Ip address 192.168.x.x

If just 192.168.x.x is entered in a browser, then it will be translated to either 192.168.x.x:80, if the browser is set to prioritize http connections, or 192.168.x.x:443, if the browser is set to prioritize https.

If just 192.168.x.x is entered in a browser, then it will be translated to either 192.168.x.x:80…

Off topic, but thanks for the lightbulb moment here! It’s obvious in hindsight, but quite an “ah ha” to realize this!

Well that was a very “simplified” answer, apparently you don’t know it first have to pass the OS and find a Matching MAC-Address it can send the request to, if it hasn’t any in the ARP table, it Broadcast ARP, and HOPEFULLY get an answer back, from HA-OS
The Browser Don’t connect to an IP, it send it’s request through your system , Browser>OS>NIC which needs a MAC-Address to send it’s packets to

A simplified answer yes, but yours are no better.
The browser connects to an IP address/hostname.
The browser never sees the MAC address.
The MAC address is used on a much lower level in the network stack and is something the OS deals with.
The network stack is made so each layers downwards can communicate with itself, which means. That means the TCP layer can communicate directly with another TCP layer and an IP layer with an IP layer.

I don’t know if we have a “communication break down” here, or we are talking from “different” perspective

A Browser communicate with The Application Layer, And Only that , It sends a HTTP/HTTPS/FTP etc. request And it’s “Out of the Story” so to speak, it Request/Receive from Layers below

And yes the request have to go through EACH layers to get anywhere, as it in the other end (the Server ) also has to go through the stack

A Browser can check it’s own cache, but have to Request i.e the OS or Routers cache, as well as it only Request a “connection” It Doesn’t handle/create a TCP/IP connection, a Browser does not i.e perform ARP-Broadcast
( A Browser doesn’t handle the TCP/IP Layer )
A Browser can however block “responses” from the Layers below, as a i.e web/Application-Server can refuse “requests” on various web-server-layers , even the OS which the Server is running on can “reject”
However a Application-server is “more advanced” Ofcause

And the “various” communication (handshake/control-functions), obviously also occur on various levels (Out of reach for an Application(Browser)

The most common cause why the Users get that “generic” Error in a Browser ( And have to look deeper into logfiles, for a cause ) , The Browser Have No idea, unless a “message” is forwarded to it, stating the cause of i.e a time-out/rejection

Data Encapsulation and the TCP/IP Protocol Stack (System Administration Guide: IP Services).

TCP/IP vs OSI
TCP/IP Model vs. OSI Model | Similarities and Differences | Fortinet.

Pick whatever you think Your Client and HA’s System Uses/Prefers

So where are we at the moment?

I’d recommend:

  1. find a PC or laptop or smartphone or tablet that is with the 192.168.68.x IP range. Temporarily is fine; we are running tests here, after all.
  2. use the browser that PC or laptop or phone or tablet to connect to these 2 link:
    http://192.168.68.120:8123
    http://192.168.68.120:4357
  3. If either of those is not working, share the screenshot, and find another PC/laptop/phone/tablet in the same 192.168.68.x network, and repeat #2 above
  4. Let us know, and we would then figure out what to do next.

With limited information so far, I am guessing the HAOS itself is working, and the issue being some networking stuff. You might want to be prepared to share your networking setup / config / HW / SW / topology , so that people here would be able to help.

I think we have a communication break down :slight_smile:

You know what you are talking about and maybe the short text part I quoted was just a bit vague, which I tried to clarify a bit.

If I were troubleshooting this, and the HA Observer page which is the same host was working, I’d open the developer tools menu in a Chrome-based browser, select the network tab, and open the IP based URL ending in :8123. This will eliminate any browser rendering issues and may give you a clue as to what’s going wrong. It will at the very least tell you if the initial http GET request and response are working. Just going through this exercise on my (working) development server the browser performs 126 separate requests just to render the default dashboard I have set.

1 Like

Yes, everything looks ok , beside i have “fallback: false” & " locals: -dns://192.168.x.x (x.x == gateway-ip)

And i have disabled IPv6 everywhere i can :slight_smile: , it’s a “local” home environment, If there is a need for IPv6 it’s a valid cause
If not, there is no point in “risking” i.e Microsoft prefer IPv6 for some “service/function-calls” , or other Device chose to.

And as k8gg Mention , Your Router-Settings can also be of value, or atleast if you run on Default-settings, the brand-name-version could be relevant ( I only knows my own Router, but others can “fill in” , if you provide the info )

I.E try to start by set “fallback: False”
Type:
dns options --fallback=false

And reboot your HOST

Better yet, assuming you are browsing from a windows OS, “route print” in a command prompt or poweshell would tell you if your router is even involved in the process. 95% of the time it’s not. Most people don’t have multiple subnets at home.