Don’t say you use ssl, but trying to access through http ?
Okay, here’s the net info output:
So it’s reading the right network info from the DHCP server, using the assigned address and the 22 bit mask.
When I changed the mask from 24 to 22 bits, I kept the DHCP address pool the same and didn’t add or change any addresses for twice the lease duration. I do have some assigned addresses specified on the DHCP server, but I do not have any systems asking for specified addresses. (Well, that I know of - if any system is asking for a specified address, it’s doing something I don’t know about.)
I’ve referenced both. I am using http to try to get to the main interface and it’s failing. But when I tried “ssh imladris” and “ssh root@imladris” it asks for my password. That tells me I’m connecting to the system. I can also connect to the Observer at port 4357 (do I have the name right - I get things like that mixed up). So the system is on the LAN and is receiving connections. It’s just not letting me connect to port 8123.
Also, side note, I have the ssh addon installed and I wanted to reach it via ssh. I’ve tried what I thought was the right password for root and for my username with no success. I’ve also tried following these instructions and have imported the authorized_keys file (I think I have - I get a success response when I use the os import command, but not actual message about importing the authorized_keys file). Also, when I try ssh root@imladris -p 2222, the connection is refused.
I get that there are two different ways to connect via ssh: the addon and setting it up by importing the authorized_keys file. I would think, if I have the addon installed, it should work that way and that having imported authorized_keys wouldn’t turn that off.
I didn’t talk about ssh, but ssl, as you mention you connect through Tailscale, i assumed this is through https(ssl)
But no places you mention whether you typed http://ip , or https:// ip, when trying to access ha
He’s using chrome.
Isn’t that browser switching automatically to httpS
Don’t know, i never use chrome 
Sorry - read that and, since I was trying to get ssh working, just read it as “ssh.”
Yes, using Chrome, and just double-checked. IT’s not adding the “s” on the end, just “http://imladris,” or “http://172.16.7.2”
I have checked and, as best I can see, I am doing exactly the same to reach imladris (the one with the issue) and lothlorien (the one in our barn that isn’t having an issue.)
Yes you are mentioning 2 system, not whether they are on the same ha versions, and not why you haven’t updated them to latest either
And have you tried to access using https ?, im not familiar how you access through Tailscale, but i assume maybe you have enabled SSL at some point, or is your Tailscale connection over HTTP ?
However your picture just below this statement show
OS 11.5
CORE 2024.2.2
That is not updated !
Thinking this over and I know you guys know much more about ha and networking than I do, but I’ve made a point several times:
I can reach this with ping, I can reach port 4257 without issue, and while I’m not logging in with ssh, I am connecting and being asked for my password. I just cannot connect to port 8123 - connections are being refused (that’s Chrome’s error message).
Doesn’t that pretty much rule out networking issues and indicate it’s something within ha?
Just saw this as I was about to hit “Reply” on this post. Yes, I updated them both last night. I believe it was an HA update and an HA core update. Might have been HAOS instead of HA, but I know it was an HA core.
(Oddly enough, when I tell imladris to update, lothlorien updates - at one point I was trying to be able to reach devices in the barn, on lothlorian from imladris and thought I had removed that plugin, but it seems like it might still be doing something. The buildings are over 300’ apart, outside of wifi range, so I can’t add Z-Wave in the barn without an instance of HA and the Z-Wave dongle on it down there.)
Just tried that. No joy.
Right now I haven’t tried to access it or anything with TailScale in a week or more. When I’ve used it with TailScale, the one change I’ve had to make when I’m away from the house and using TailScale is to use “http://imladris.arda.ea” instead.
Last night I had the two pending updates on both systems. I did one of the pending updates, about 10 minutes later I came back, the update was done, then I ran the 2nd update and went to bed, to wake up and find lights that should turn off after sunrise were not off.
On the other system, here’s the version info:
If it’s showing as not updated, then maybe the issue is that I updated the core and something went wrong with the OS update?
No & Maybe
Updates to latest HA( OS & CORE ) and don’t say you have when you haven’t
Sorry above makes no sense to me, maybe a language barrier
Yes latest pic is an updated version, the first pic above is not updated
Type in the console
core update
os update
su update
PS: And in regards to network “issues” , if you are not certain that you need ipv6, then disable this in your Router/DHCP settings, as you don’t want “some” devices like i.e Apple using this.
Not that i know the reason for you to “expand” your network segment, i understand it was not to expand your DCHP-Scope, so still limited to same.
If i would make a new network ( or Extend it ) , i would move my Router/DNS/Gateway ( And as mention if you haven’t got a valid reason for using IPv6 , then disable it )
Roughly how long should I have to wait on the updates? 20 minutes so far on “core update.”
I’m wondering if I might do better doing a backup to a USB drive and doing a fresh install, then importing the config into the new install. Does that add the plugins and everything else that I had on it previously? I’m thinking maybe there is a problem with the USB stick the system is on.
Regarding my comments on the update: The main point is that when I update imladris, lothlorien always updates too - automatically and without me starting that update. I was just thinking this might be due to an update I tried using so I’d have the devices on the lothlorian system also available on the imladris system. However, I thought I had deactivated that add on.
I have no idea which “add-on” you are talking about, but i guess it’s the same you call a “plugin” above ( I is 2 quite different things in HA, however i don’t know what you are talking about anyway )
However this didn’t seems to work, as you clearly have 2 different version
are you here referring to the “plugin / add-onn” and now calling it an update ?
Anyway as you haven’t “described” this functionality previously, i guess we better keep this out of the equation.
It’s also “new” info that you infact runs HA on a USB stick, attached to your PI , or PI’s ( If on both) Beside you also don’t tell whether it’s PI3b or whatever
So have patience with the update, some takes 5min, other takes +20min , depends on alot various factors ( to make a short statement )
It took a long time, but the core updated and I could get to the web control panel and the mobile app started working. I went on and did “os update,” and that worked, but “su update” didn’t work - it told me no “su” command.
I figure the plugin or add on that was supposed to link the two systems wasn’t working, which is why I left it out earlier. I mentioned it only because it’s the only reason I could think of that would explain why, when I start an update on imladris, lothlorien also updates at that time, too. It’s interesting that I started the first update, both systems updated, then I started the 2nd one and it worked on lothlorian, but did not work on imladris (the one I started the process on). So both started the update at the same time, but imladris, for some reason, just didn’t do or didn’t complete the update.
Well, it was working, seemed to be updated (using V 12.1 instead of 11.5), but now it’s not working again.
So what about backing up, from the command line, and making a new install and restoring? What do I lose if I do that? Do I have to install plugins and add ones again, or will those be automatically redone?
Try the full command ha supervisor update
I would remove any integration or add-on that can alter access to the basic installation. From remote access, to special CLI
I say that because “it works for awhile then stops” which points to either a integration or add-on that takes awhile to load and begins it’s control of the HA.
Also I see this is run inside of a docker. Is it inside a container, inside of docker?
There are many articles on the web. I used this to find them: “how to run home assistant inside of docker”
I removed the one integration I was worried about. I don’t know if that was it, but it seems to be working now - other than ZWave JS UI, and I think that’s because I had to plug in a USB hub so I could have a keyboard and extra USB stick in there at the same time so I could add authorized_keys for ssh. (Turns out I finally got it working through the ssh addon, instead.) I’m down in the barn for a while and I suspect when I get back to the house and can pull the extra USB devices, that ZWave will work again.
I use the Raspberry Pi version of HA, so it’s the default Pi install. I never realized that used Docker until today.
I did see, on the web UI, an update pending for HACS, but when I tried to install that update, it failed and now I can’t see it as an install option anymore So, once again, I’m thinking about the idea of backing it all up and making a new install. Very likely I’ll try that now, since, if it fails, I can go back to the current version easily. There’s a long story why I’ve been using USB sticks for HA in the past, but this would also give me the chance to move it to a microSD card.
I just haven’t looked into transferring a configuration before and don’t know how much transfers - if addons and integrations are automatically reinstalled and so on. I’ll look into that in the next few hours or tomorrow.
I read up on backups. While I’ve migrated from one system to another in the past, it’s been a good while and it’s really hard to keep up with what systems completely restore from a backup and which ones still need you to add on plugins or addons manually and so on. Once I read up and saw that backups were complete and that people had no problem transferring their data from one system to another newer one by importing a backup, I was seriously relieved.
I used Raspberry Pi Imager to put HA for a Pi 4.0 on a microSD card, then made a backup to a network shared volume I had added in Settings->System->Storage. I was still hoping I wouldn’t have to do this, but the ZWave system wasn’t working. (Odd - in ZWave JS UI, it could read all the ZWave devices in the house, but the actual integration wasn’t working.) I shut down, powered down, removed the USB hub with some extra devices in it, and tried rebooting. It took a long time due to fsck dealing with the volume on the attached USB stick. The system couldn’t do a full reboot and dropped into the emergency CLI instead.
I powered down, put the microSD card in and removed the USB stick and booted from the memory card. (I always change my Pis to boot from any inserted USB stick before booting from the microSD card - it’s come in handy many times.) When it was ready, I imported the backup I made less than an hour before.
Everything restored 100% and everything is working now. There are no messages of pending updates, either. My theory is that the USB stick was having issues and that the OS tried to update and could not (so it was at 11.5 instead of 12.0 or 12.1) because of problems with the integrity of the USB stick.
TL;DR on USB and microSD cards: My wife and I built a house out in the country a while back and moved out here to deal with crappy internet from ViaSat and power flickers. While we have fewer blackouts here than at my old house, in the suburbs of the state capitol, we do get power flickers. It took me a while to get UPS devices on all my electronic devices. We have a whole-house generator that will keep power on for the house and barn, but it takes 45-50 seconds to kick in and provide power. So with power flickers, and me still trying to get UPS devices for every outlet in the house and barn with any computer or computer like device on it, I had some Pis that had to deal with power flickers. (By flicker, I mean on-off-on-off several times, but never off for more than a second, at most.)
One system, in particular, had issues. After every power flicker, it would stay down. I’d take the memory stick from it, in the barn, up to the house and check it. After several times dealing with this, I found that just reading the microSD card on a Linux system that could read the fs on it, things worked fine. So something in the filesystem was getting hurt when the Pi lost power (or had power flickers). So I replaced the microSD cards on those Pis with USB sticks and continued to use USB sticks until I had time to do some research and find out what microSD cards were made to withstand issues like that (as opposed to ones intended to just go in cameras or other devices and not deal with anything like a power flicker).
That was a few years ago and my HA systems haven’t been perfect, but they’ve kept working to the point it would be more of a pain to back up, make a new image, and restore the backup. So it’s taken me time to move some systems from USB sticks to microSD cards. I know USB sticks are a LOT slower to write to, but are less volatile in some situations, but from what I understand, most are not designed to serve as an fs on a system for years. I’m betting the USB stick was just wearing out and I got lucky and got a backup from it before it died.
So now, tomorrow, when I have time, I’ll be making another image and upgrading my other HA instance from USB stick to microSD card.