I have rebooted several times and it doesn’t come up.
(Side note: The help for the command line is confusing. When I type “help ha os,” one command it lists is “restart,” but when I type “ha os restart,” I get an error message. Just what do I type to reboot the system from the command line?)
In chrome, try http://172.16.7.2:4357 - that will show you the observer page to see if everything’s up and running.
If everything’s ok, try to reach it with the IP address directly http://172.16.7.2:8123 (change that to https and click past the warning if you have external access enabled).
If you can get in, it’s either an issue with mdns or an ssl error, similar to this post.
The one with the orange square in the bottom is an iOS error.
Make sure that you are not using a VPN and that your WiFi is not segregated/segmented in the AP/router.
I also tried both with the system name and the IP address from my phone, via wifi, with the same result. (My desktop is connected by CAT5e, so no wifi involved there.)
Since Chrome is saying, “Connection refused,” I thought I’d try ssh, but I realized my HA interface username doesn’t seem to be acceptable for ssh.
Currently no VPN on the network. Also worth noting I can get to the other system from my desktop (on ethernet) and my phone, on wifi, with no issue, but can’t get to this one either way.
I forget the term, but my wifi passes the DNS requests through to my pfSense DHCP server. (Or is it just “pass through DNS?”)
I’ve always found the console frustrating - like I mentioned above. I type “help” and get commands and, for example, when I see “system” is one, I try “help ha system” or “help system” and get restart as a system command, but “ha system restart” or “ha restart” don’t work. I’ve had that frustration with other commands.
Good thought - and it’s my bad I didn’t include those in a screenshot. Just general troubleshooting advice:
I’m not using a proxy server or VPN and I know there’s a connection because I get to it to see the Observer and I can ping it and it responds to ssh, even if I’m not sure which username to use for ssh.
This makes me think it’s not an issue with my LAN or the wiring or wifi (especially since no other systems, including the 2nd HA instance, are offline) and that something is going on in HA itself, some kind of security setting in HA or the system that causes it to block a login. It was fine until last night when I did an update for HA and HA core.
Yes. I’ve been using names for a few years. The HA in my house is “imladris” and the one in the barn is “lothlorien.” I don’t use “.local,” I use “.arda.ea” and I’ve never had an issue with any system not being easily reachable by name - except when I’m outside and use Tailscale. (I’m on an ISP that uses CGNAT, so I can’t do port forwarding.) With Tailscale, I have to use something like “imladris.arda.ea” to reach that instance of Home Assistant. Note I’m having the same issue whether using the system name or the IP address.
Regarding the DNS and DHCP potential issues: I’ve been using 172.16.7.0/24 and have switched to 172.16.7.0/22 a month or more ago and I have a few systems in the 172.16.6.xxx range, but that shouldn’t be a problem with this one (172.16.7.2) - and, remember, this only became inaccessible after an update. It worked fine before that update and the DHCP server address space change was over a month ago.
One thing which is relevant, is readin the text, when typing " -help or -h ", you would find the right syntax then, i.e "-help AFTER a command , i.e there is not such thing as “system” but OS, Core & Supervisor
i.e
Help me out on this - I’m not at all clear about what you mean by “overlap.” Are you saying that this address (172.16.7.2) would show up more than once? I do get that it now includes a range from 172.16.4.x to 172.16.7.x. At this point I have only 2 machines in the 172.16.6.x range, the rest are all in the 172.16.7.x range - including the machine in question.
Okay, so don’t use just “help” to get the more detailed commands. And it’s possible, since the HA system is downstairs, that I was thinking “system” instead of “os.” But I get the important issue - don’t type “help ha os” or something like hat, use “ha os -h.”
I will probably also go check on what I need to do to set up ssh so I can reach it that way and I don’t have to run upstairs and downstairs to try different things. (I did find a good thread on what to do to enable ssh access.)
HA can contact all machines on 172.16.7.x without needing to route through 172.16.7.1
Device assumes it can contact all machines on 172.16.7.x without needing to route through 172.16.7.1
Watch what happens by just changing the mask: /22
IP Address:
172.16.7.0
Network Address:
172.16.4.0
Usable Host IP Range:
172.16.4.1 - 172.16.7.254
Broadcast Address:
172.16.7.255
Total Number of Hosts:
1,024
Number of Usable Hosts:
1,022
Subnet Mask:
255.255.252.0
Wildcard Mask:
0.0.3.255
device assumes it can contact all machines on 172.16.4.1 - 172.16.7.254 without needing to route through? - I dont see your router address, assuming the IP ranges overlap I’d bet you didn’t change it so for this exercise - still at x.x.7.1
We’ll start here. What is the IPv4 address, netmask and router for the CLIENT that is throwing the error?
(We need that to solve this)
What is the ip address of the router being assigned
The router is a pfSense system at 172.16.7.1. It’s also the DNS and DHCP server.
My desktop is at 7.7 and the mask is 255.255.252.0. I can reach my systems in the old range of 7.xxx and the two systems in 6.xxx from it, using both ping and using my browser. (The two in the 6.xxx range are running OctoPrint.) When I tried ssh to reach the HA system, it did connect and ask for my password - I stopped there since, at the time, I wasn’t sure what username or password HA would accept. So I can reach the system in doubt from my phone on wifi. (I have a terminal app on it.) I can also reach it from my desktop with ping and reach the HA Observer at port 4357 on my desktop with Chrome.
@NathanCu is explaining it more detail, then I went into. But you widen the range in the router which is okay if running out of addresses for use. But you widen the range for the HA to use also. That is the overlap and possible conflict is coming from.
Okay, I get that now. So, since I have my two Pis running OctoPrint defined in HA, and those two are in the 6.xxx range instead of teh 7.xxx range, HA might be having an issue there? And maybe I need to change the netmask on HA or something like that?
Exactly. Devices that obtain ip addresses. In the new expanded range may have a net ask conflict with devices in the old space. (especially during the dhcp lease window but also any preexisting manually addressed client that doesn’t get the net ask updated.
In short - some clients get confused and start sending traffic to the router which drops it because ‘you don’t need to route that dum dum’
So even if I tell systems to use my one system as gateway, DHCP, and DNS, and to get info from there, they might have their own netmask set that might make it hard for them to access some systems unless I change that info on that machine?
(Right now I’m using this page to get ssh access so I don’t have to keep running up and down the stairs between the HA system and my dekstop.)
Any that are DHCP will change when they re-lease. The network would have been REALLY dicey for about half of the time of your standard DHCP lease right after the change. If you have a standard 8 hour lease or even a 24 hour one you should have long cleared it. But if you’re setting manual addresses in your router or on the machines themselves, who knows…
