#installation-archived
1 messages · Page 42 of 1
if not raspberry pi with HAOS, if not my APU2 with HAOS, whats the next best? im confused about all the different pieces that make up HA and if something is missing, i might miss some functionality
vm with HAOS ?
Hey everybody! All of a sudden I can’t access my server at all from anywhere. Nabu casa can’t find the instance… neither can my iOS OR OSX app… and I cant even find it when I type in the server’s url — I’m so frustrated I think I’m just gonna wipe everything and start over. Any words of wisdom as to best practices of how to do that?
I have a dell 3050 optiplex that I installed the HAOS onto directly
I did the generic setup with the “write haos to boot medium” thing…
Connect to console? Meaning do i have a monitor setup to the server? I do and everything looks normal! So frustrating
well if you ahve a keyboard and monitor
are you able to see if the network info is correct?
Yes. And it is
Wait do you mean to confirm that the internal url is the same one im typing in?
what does the net info command say
you see connected: true and supervisor_internet: true?
And does it have a new IP address from your router vs the one you're using to connect?
if connected: false is there, check your network cable
Supervisor internet is true
do you have more than one interface?
I use the desktop app on Mac, and my iPhone app
I mean on the dell 3050
under net info
And I had just signed up for Nabu Casa when I stopped getting access
Does the HA computer have wifi and ethernet?
you have to use the nabu casa link then
Yeah it’s connected via Ethernet
It says enp1so
I feel like a screenshot/picture makes more sense
Oh wait it says: “w1p2s0”
did you try the nabu casa link to get connected?
it really seems like you're on wifi rn if one adapter says disconnected but the supervisor thinks it has internet
Hmm. How do I turn the wifi off?
well do you have a network cable plugged in?
did you try this?
Yeah. Just says “connection failed: no response from instance”
well from the console you can type login and that should bring you to a # prompt where you can run nmcli radio all off
that will turn off the wifi
Sorry im a total dummy with these commands… so just type “login” and enter?
Did it. There’s another # below it now
now what does net info say
Are you 100% sure both ends of the ethernet cable are properly plugged in by the way? It sounds silly, but it's an easy mistake to make
Connected: false enabled: false
supervisor_internet: ?
Lemme check
True!
What#s happening?? Haha ughhh
Would a reboot help?
Is there a space between the g and the 1?
yes
Says: 1.1.1.1 is alive!
Yes connection is solid
are you using SSL by chance?
If I am it’s by accident
I do remember messing with the https vs http settings on my computer before my server became inaccessible
Is it possible that I messed something up??
when you say the http vs https setting do you mean in HA?
Honestly I don’t remember. I was setting up Nabu casa at the time
Probably not on the server tho. I don’t have enough know how to change that
do you have backups from before the nabu casa change?
it's ok, I too am an idiot
Should I just refladh my servers hd?
But youre also SO KIND to be helping me like this
I've locked myself out numerous times in the last 4 years or so
Haha im about a year in and this is my first time. I was proud of how much I had learned… until this lol
I mean you can troubleshoot it and maybe get it working
Nah I’d rather just start fresh honestly
or... spend like 1 hour, have a beer while it restores
Yeah exactly
Last question
Is there any reason I can’t just pull the hd and reflash it?
Or do I need to wipe it first or anything?
then dl the latest one to your pc, and restore it
yup
Boom. Will do. Problem solved
KrAzle, you are the BEST!!!!!
I can’t thank you enough
don't say that, others will see
thank you is more than enough 😄
Seriously. I appreciate this so much.
Happy to help lol
thank you 😛
@brazen herald thanks for your help too!!
Any fun automations?
define fun, process not found
well it's just to turn my hallway lights on when I get home after dark
Not so much “fun” as “helpful” I suppose
if I was lazy I would set them to come on at dusk like the outside lights but I'm only a procrastinator, not lazy
Hahhaa. Gotta do it right… at some point
That should be the motto on your family crest
actually the family motto on the heraldry is "initiated, dedicated "
I should add "but just not now"
is there a light weight install of HAOS for ubuntu? id like to use frigate and zwave and other add-ons.
I currently have a pi4 with HAOS but want to add frigate, so i think i need to move the install to my nuc
HAOS is an operating system. There is no HAOS for Ubuntu.
unless you run it in a VM
But there is no "lightweight HAOS install", there's just the official one
what kind of VM software works? and is it easy to manage? i guess thats what i want in being "lightweight"
pi4 was easy, install and never have to do anything but click update
maybe i should just use frigate on another machine and keep the pi4 as is
What you are after is docker
There is also a Frigate add-on - endless options
ok it seems that frigate in docker on nuc and HAOS on pi4 will work
A lot of people have good experience running HAOS in a VM on Proxmox.
I'm not sure I'd run frigate as an HA addon. Maybe cleaner / more reusable as a separate container or VM.
(if it can even be an addon, that is)
It can be but I know I've seen a few people recommend not running it that way.
You can run haos as a vm in kvm on ubuntu. or you can run home assistant as a docker container, and have frigate zwave etc as other docker containers, on ubuntu.
i can not connect my phone app to home assistant
well that doesn't answer any of my questions
i dont have any acess
and that still doesn't answer my questions
a
b
be explicit when explaining what you tried. I.e. post the urls you've tried.
i am a new user i followed all the steps to install my home assistant but when i tried to connect home assistant with my phone using home assistant app , all url entries were invalid and not accepted
c
dude
Answer my questions
being a new user doesn't mean you can "not read"
so read my questions, and answer them
do not avoid them like you have been
no
If your phone is local, then your local address will be http://homeassistant.local:8123 or http://<yourip>:8123
was that really that hard?
no, and stop using caps
do you have a screenshot of that url w/ ip in the application you're connecting with?
use imgur or something similar to post images here
yes in french
please post the screenshot then
i can not send screenshot, apparently i need permissions??
but it is writen
impossible to connect to home assistant
lets go back a few messages and read what I said again
what does this say right here?
Hmm, i wonder
Where did you install Home Assistant and what does it show if you connect a screen to the device you installed on?
Ok, so you're using andriod, so you should check out the #android-archived for further help. It seems like a network issue because everything you're showing on screen appears to be correct assuming you're using the correct IP and you're connected to an SSID that has access to your home network.
the app on my phone detect the url adress http://(my IP):8123 but i can not connect with
No not the App, where is your home assistant server installed?
thank you very much for your help, i am usnig android and i tried also with my ipad but i counter the same probleme
how do you normally connect to HA?
i used rasberry pi imager and followed all the steps up to the url adress where i could not access
So attach a screen to the PI and see what it shows
It is possible the PI has not finished setting up yet, or the image was not correct.
i never connected to HA !!!
ok, so how do you know if the IP address is correct or if it's even done setting up?
do you have a computer?
If it was a Pi 3 it will take ages.
According to the led lights on the HA the installation was successful
because the app detcted the adress but i could not connect
this was my adress
It might still be installing. Put a screen on the Pi or try it again in an hour, if you can not be bothered to plug it in.
You should connect to your router and verify that it's at that IP address
secondly, you should use a computer during initial setup or just a normal browser
not the app
So open chrome on your phone, and go to http://192.168.1.44:8123/
it is a pi4
i tried with my laptop using http://192.168.1.44:8123/ but did not succeed
same probleme as my phone app
what do you see at that URL?
i did many installations
Already done but not working
You need to connect to your network
That image is telling you that you aren’t connected to a network
i am connected to the network
You need to plug in a network cable on the PI it will not use Wifi
At least not initally
i am using it to send you messages on discord
with the same laptop but can not acces to my HA using http://192.168.1.44:8123/
yes HA is not at that IP
Ok, then you have to find out what IP it’s at by looking at your router
Usually that’s 192.168.1.1 but it varies depending on what router you own
did not find it !!
Then your HA is not connected to your network
Do you have an Ethernet cable connected to your pi and router?
yes of course a cable is connected between my HA and router
Ok, then is an SD card inserted properly into your PI?
and is the SD card flashed with HA?
Also, is the PI powered on?
i used the sd card for the pi image , it is not inserted now
the pi is powred
and led lights flashing normaly
you need to have the SD card in the PI...
turn off the pi, put the SD card in, and turn it on
it's going to take a bit to setup, so be patient
I'd walk away for an hour and then try to log in via your desktop browser
no problem
I have home assistant yellow with a selected pi4 8GB ram anD 32GB, Maybe i went on a wrong pi image !!!
If you have a slow internet connection, it can take hours to set up
You either have a HA Yellow (no SD card slot - wrong install medium) or a Pi4 (if you used the Yellow image for that - it's wrong)
hmmm whats a normal amount of time for a restore to take 😬
haha thats like 2 days for me 😄
is there a way to check the restore progress by any chance?
Not that I'm aware of
how big is your database?
The restore file was like 50mb.. but hear me out because this may have been a bad idea.
I did a backup yesterday, then took the drive out and put an old drive with a HA install from 6 months ago in, then ran the restore from the new file..
The silence is concerning. Haha.
50mb is nothing
my backups are 1 gig IIRC
although I never run a normal backup
er restore
I just extract the files manually
mine are 730mb
its been going for 30+ mins now, just stuck on connection lost
look in your logs
cant, connection lost
ssh or samba share...
HA 101, give yourself access to the system when HA is not running
i.e. ssh or samba share
for some reason I cant get in the config folder :/
SSH will allow you to see the supervisor logs
but I can get in media/share/ssl/etc
do you have ssh running?
to be honest - I'm not sure, I don't use it but its possible
downloaded putty, put in the ip, port 22, and it fails with network error
so perhaps not
then you might have to just wait
or start over witha fresh install and then do a restore
okay, thanks for the help 🙂
by some miracle I just turned it off with the power button, turned it back on, and it worked and had restored correctly..
Hey-hey i am still stuck at core 6.2 what can I do to update? I tried from ui and from terminal but nothing happends. Can you help me debug the issue?:)
do you mean 2023.6.2? if so, what is your installation method?
Yes
I tried a few method:
- from ui by clicking on the update button - from terminal add on by typing ha core update, and also teied —version 6.3 so update it only one version - i tried the same from ssh too just to make sure
sounds like it. So for the update, you tried ha core update --version 2023.8.3?
and if it doesn't work, it's probably DNS or a space issue
or it's Supervised with a broken installation
Yes it is haos, running on an rpi
Yes also tried that
I have 400gb free so I dont think it is a space issue
If it is dns how do I solve it?
Check the supervisor logs (select Supervisor at the top right): https://my.home-assistant.io/redirect/supervisor_logs/
You can also run ha supervisor logs via SSH if you prefer that
What should I look for?
something broken
Okay cleared the log, pushed the update button.
This can cause an issue like a described?
Logger: homeassistant.components.websocket_api.http.connection
Source: components/websocket_api/http.py:53
Integration: Home Assistant WebSocket API (documentation, issues)
First occurred: 20:16:58 (2 occurrences)
Last logged: 20:17:39
[547683556176] Disconnected: Did not receive auth message within 10 seconds
[547713479888] Disconnected: Did not receive auth message within 10 seconds
I have HAOS installed through Proxmox on a mini office PC. I thought I'd add other nodes but here I am a year later and still do everything through HAOS. I've been running into storage issues as I add media for my Plex server to the small SSD in the machine so I started thinking about other options. I'm pretty sure I could mount an external HDD for media to the PC through Proxmox, but I think it's only adding an extra level of complexity since I don't use Proxmox for anything else. If I wipe the drive and start again with just HAOS, is it as simple as plugging in the USB HDD to access my media? Ideally I'd have the internal SSD for just HAOS, and then the external HDD plugged into the machine for media.
you won't be able to use the plugged in drive
network mounts are the way.
install the harddrive in proxmox, share it, mount it in HAOS
when it says it may take 20 minutes, on a pi 3, it is no joke
no, that's the joke.... it will take much longer than that
couldn't get the networking right in kvm/host interfaces, so I thought id get one of the pi 3's out of storage.... ooof
Grab a snickers
40 minutes, and i grabed a beer
Hi I'm trying to install the Esphome into my home assistant, in google / esphome webpage etc their guides are different. They have add-ons where as i dont have that in my docker container run home assistant instance. The only way I could find to install esp home is through integrations, and it asks an ip address and a port ? Am I supposed to put home assistants ip address? since thats where i'm trying to install the esp home ?
hmm even if i do that it says :Can't connect to ESP. Please make sure your YAML file contains an 'api:' line. ....
The container method has no addons: https://www.home-assistant.io/installation/#compare-installation-methods
You have to install ESPHome as a separate container yourself: https://esphome.io/guides/getting_started_command_line.html
The integration is glue between them, so to speak
ok then me goes to install that docker 🙂 will be back 🙂
oops, just remembered my home assistant is actually running on host to be able to reach all the devices in the house network. Now that i will have to add an ESPhome to that stack, it will also be on host. Does ESPhome container need network access or is it just to aid home assistant ? does it only need to speak with Ha ?
hmm it even wants prievilaged
The docs actually recommend host mode. See second command here: https://esphome.io/guides/getting_started_command_line.html#bonus-esphome-dashboard
Awesome, just a follow up question ; I got it working. (ESPhome in docker at host, and accessed it via its exposed port web interface. Now i realised that we can actually flash devices straight from the web browser ! Anywhere in home any device! Thats so good. But IT have a specific note saying : Connect the ESP32 or ESP8266 to your computer and click on connect to start managing your device. | Does this mean that i can't flash my nucleo 144 stm32 boards ? Or is it going to be a trouble if i do so : (
Just trying to understand if I should wait a few days more to get my ordered ESP devices rather than messing this nucleo 144 ?
If you're using docker run I recommend swapping to docker compose, a lot easier
I'm using portainer, for now i'm ok with it. I can write those yaml files and mostly it works. Also its a synology server so for docker compose just dont wanna leave ssh open and log in every time
But you don't have to log in every time? You run docker compose up -d
And the stuff keeps running, same as docker run
Unless you mean going into edit the files, in which case, doesn't really matter
🙂 Ok I will learn more docker compose but thats for another weekend haha 🙂 Just curious i asked Impact but hes probably busy, how do we install HACS on home assistant where we run HA from a docker container ? Previously he advised to install ESPhome from docker container which worked. But now i want that local tuya HACS integration
Do I need to install a docker called HA HACS or something like in a similar fashion ?
By following the installation instructions. There's a container section
oops I found it 🙂 Just needed to authorise my github ! thanks
I actually answered before you asked here 🤔
sorry, i must have not seen it , was dashing through my weekend adventure 🙂
thanks for the advices!
anyone having any suggestions for issues with hassio_dns, im running home assistant os and have been trying to get it worked out. I was moving my installation from a vm to a sff workstation. I tried multiple times to start it using docker start but it keeps shutting down.
Check its logs via docker logs ...
Thanks, didnt think of that command. Waiting for diskpart to wipe the disk and do a fresh installation without anything to further diagnose.
Anyone can help me out with this? What should I look for or what am I missing?
Someone mentiones dns issue so I enabled fallback dns from cli let see does it help.
After my SSD failed, I installed a new one in my NUC. I installed HA and then restored the backup.
I've restarted HA several times (full restart). However, I still have several devices that can't be loaded properly. gardena_smart_system, pyscript, tesla_custom, bosch, easee, miele cannot be loaded. Some other devices can't be initialized correctly, e.g., solar Forecast, google Nest. Has anyone had the same problem or any ideas?
Other devices and integration work fine (for example ZHA).
Has anyone had any ideas or experienced the same problem?
hi. I'm having a lot of issues with my HA installation in my Rpi4. I can access the web interface but I can't access a lot of options. At the moment I just want to try to create a backup before I format everything but I can't access that option in the web interface. I get the error that I need to restart supervisor. When I first restart the pi, if I go to homeassistant.local:4357, I can see "Supervisor: Connected, Supported and Healthy". On top of this, it seems that my ssh port is not accessible and I don't have a micro hdmi port. Is there anything that I can do to enter the command line in any other way?
You'll have to post logs explaining the failures to determine what's going on
Protokolldetails (FEHLER)
Logger: homeassistant.setup
Source: setup.py:209
First occurred: 13:37:43 (5 occurrences)
Last logged: 13:55:13
Setup failed for custom integration tesla_custom: Requirements for tesla_custom not found: ['teslajsonpy==3.9.2'].
Setup failed for custom integration gardena_smart_system: Requirements for gardena_smart_system not found: ['py-smart-gardena==1.3.7'].
Setup failed for custom integration pyscript: Requirements for pyscript not found: ['croniter==1.3.8'].
Setup failed for custom integration easee: Requirements for easee not found: ['pyeasee==0.7.53'].
Setup failed for custom integration bosch: Requirements for bosch not found: ['bosch-thermostat-client==0.23.2'].
Would other logs be more helpful in order to find the solution?
What release of HA do you have installed? The Bosch one for example needs 2023.7 as a minimum.
Hello I have a Home Assitant Blue and I am just now getting around to setting it up. I get an error in the supervisor log about the docker image can someone help me please
is there a way to flash it with the latest OS
I am sure it is out of date
Same way you installed it
I bought the Blue with it pre installed I thought
Oh. Not familiar with that but maybe this works: https://www.home-assistant.io/installation/odroid
That works, you just need to remove the Blue's case. (I think an Allen or Torx key is needed)
what file should I download to flash it with
It is stated in the documentation
Hello everyone ! I am starting to lose hope and would like some help. Quick link to the explanation of my issues: https://community.home-assistant.io/t/need-help-many-issues-probably-related-to-each-other/604750. I would be sooo grateful if someone can help me with that!
Sounds like a networking and/or DNS problem in your setting
./dnshaiku
ZHA, so you can use the HomeKit #integrations-archived to expose it from HA to HomeKit
Not that any of that is installation related
was this message for me ?
And this one ?
But yes - i am pretty sure this is all linked to a DNS / firewall issue
the strange thing is that the setup is the same as my other HA which works just fine
Any thing I should check specifically ?
Hi , I have a problem, my home assistant server is on a raspberry pi 3 and when I start it ,it restarts in a loop.
Can you help me please ?
Can you connect a monitor and check what it displays?
Yes
Please use imgur or other image sharing web sites, and share the link here.
Image posting is blocked in most channels to discourage people from sharing text as images. Sharing text as images assumes that everybody sees the world as you do, which isn't the case. Some people are colour blind, or have visual impairment that means they can't make sense of an image of text.
I've installed the latest and updated: Home Assistant 2023.8.3 . I think I also have a DNS issue. all Local integrations work fine but any which need internet access are brocken and any features which need internet access have not functionality as well
nslookup times out
You get bonus points for it being horribly out of focus 😛
Yeah I can't read 💩
bad news is shit's fucked (technical term) most likely
good news is you can restore from backup to another sd card.. but that assumes you have backups downloaded off the device
which you do, right?
How i can know if i have backups downloaded off the device
Well.... if you have to ask the answer isn't likely to be good
Did you ever set up anything? Any add-on to do that for you?
On home assistant I checked the box to make a backup before the updates but that's all I think

do you know what is the source of the problem?
Backups don't help, if the system they are on dies
It's possible only some parts of the data is corrupt. Worth a try to plug the card into another machine and check
And how i recover uncorrupted parts?
I'd plug the SD card in a card reader and plug that into the PC. You need something that can work with ext4 to read it though
Ok thank's and what did I do wrong to get this? So that I don't repeat the error
You need backups on a different device. NAS or google drive would be one way
Pi's eat SD cards
What you did wrong was to use one
An endurance or dashcam card lasts longer
An SSD (on a powered USB hub) lasts even longer 😉
I've drilled down my issue after the new install (new SSD) and after the backup to an DNS issue ➜ /config ping 8.8.8.8
PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data.
64 bytes from 8.8.8.8: icmp_seq=1 ttl=114 time=36.3 ms
64 bytes from 8.8.8.8: icmp_seq=2 ttl=114 time=30.7 ms
64 bytes from 8.8.8.8: icmp_seq=3 ttl=114 time=42.1 ms
64 bytes from 8.8.8.8: icmp_seq=4 ttl=114 time=27.8 ms
64 bytes from 8.8.8.8: icmp_seq=5 ttl=114 time=23.8 ms
^C
--- 8.8.8.8 ping statistics ---
5 packets transmitted, 5 received, 0% packet loss, time 4004ms
rtt min/avg/max/mdev = 23.752/32.128/42.057/6.425 ms
➜ /config ping www.google.com
ping: www.google.com: Temporary failure in name resolution
➜ /config
It has always been
ever since in-addr.arpa
any pointers?
Can you explain why disabling the fallback would solve the issue? "ha dns options --fallback=false"
I'm trying to understand the issue but neither the website nor on the forum I could find proper information
And I still do not understand why this issue popup after restoring a full backup
Did you have something like AdGuard Home on your HA and everything was pointed to it?
nope. Duck DNS and NGINX
Is there something missing from the backup restore process I'Ve overlooked?
It seems other long time users also have dns issues after restoring from a backup: https://community.home-assistant.io/t/flashing-hard-drive-get-stuck-at-31-balena-fetcher/604336/8
all from the current week. Is there something going on?
Have you rebooted the host after restoring the backup?
Hello every1, want to order the HA Yellow, but saw that it doens't have any CM4 module, for sure that needs to be bouth separate, also do i need to buy a ventilator? and how does the software is installed or is already installed even without the CM4 module? 10x in advance
For the variants without CM4 (called Kit) you need to buy one yourself. No fan needed. The software is usually installed with an installer USB stick you need to create. That writes the OS to the eMMC of the CM4 (if it has some) or the NVMe (for lite variants and when forced to do so). Nothing is installed without the CM4.
Buy something better
If you buy the variant with preinstalled CM4 (the 2GB RAM, 16GB eMMC variant - called Standard), the OS will be preinstalled
like what else?
https://www.ebay.com/itm/385670237223 link i've shared previously for an $80 mini pc "m910q" there are many variants on ebay for reasonable prices with 8gb ram or higher and 128gb ssd/nvme or bigger
on the site all of them are without CM4, but found this one locally https://www.optimusdigital.ro/ro/placi-raspberry-pi/11524-raspberry-pi-4-model-b-8gb-765756931199.html
That is no HA Yellow
no but is the missing CM4 that i need to install, right?
ok, can you please share exactly with it needs for the Yellow?
Something like this. But there are tons of variants to choose from - when they are available.
https://www.optimusdigital.ro/en/raspberry-pi-boards/11769-raspberry-pi-cm4-4gb-ram-32gb-emmc-memory-wifi-pcbext-4104032.html
so... it would be better to have the computer than the yellow... keeping in mind that CM4 is shortage, right?
As much as I personally like my Yellow, you get much more performance with a mini PC
And they are widely available
but with the mini pc you need the zigbee antena? https://sonoff.tech/product/gateway-and-sensors/sonoff-zigbee-3-0-usb-dongle-plus-p/
That's correct
Or an off-pc coordinator
about 20 times already; Full Restart (System(
Hi, I have a question, to install home assistant is it better to do it on a rasbpberry or an old PC ?
Hey @pulsar timber same topic that I had last night, using a mini PC is the best option at the moment, as the CM4 are expensive and hard to find
Crashed no, after 7 days! I can ping the Raspi, but won't be able to ssh into HA.
And if i have a raspberry pi 3 which one do I choose?
Hi, I want HA for my smart home and want to know what’s the best hardware
#hardware-archived can help
@placid sage I converted your message into a file since it's above 15 lines :+1:
You need to be on Python 3.11(.4) to update HA past 2023.7
You'll want to install that and try again with a fresh venv
Even if I try to update to any higher Home Assistant version it fails. I tried different versions lower than 2023.7
Installing Home Assistant on Proxmox question, maybe simple.
Any Chance to update to Version 2023.7 or lower with python 3.10?
Why not just upgrade to python 3.11
any chance that this is what causes my issue also? I am stuck at 2023.6.2, so no matter how I try to upgrade from cli or UI, tried with dns fallback enabled also.
I am using HAOS so probably not, just to make sure 🙂
did you try with it disabled?
Well... clearly not
Hello, I'm stuck at preparing for an hour haos_ova-10.5.qcow2 Observer shows everything green. I'm using openmediavault if it matters
it's a hurry up and wait game
previously i tried 10.3 and after entire night i changed to 10.5
[supervisor.docker.interface] Can't install ghcr.io/home-assistant/qemux86-64-homeassistant:2023.8.3: 500 Server Error for http+docker://localhost/v1.42/images/create?tag=2023.8.3&fromImage=ghcr.io%2Fhome-assistant%2Fqemux86-64-homeassistant&platform=linux%2Famd64: Internal Server Error ("Get "https://ghcr.io/v2/": dial tcp: lookup ghcr.io: no such host")
23-08-21 10:29:20 WARNING (MainThread) [supervisor.homeassistant.core] Error on Home Assistant installation. Retry in 30sec
23-08-21 10:29:50 INFO (MainThread) [supervisor.docker.interface] Updating image ghcr.io/home-assistant/qemux86-64-homeassistant:landingpage to ghcr.io/home-assistant/qemux86-64-homeassistant:2023.8.3
But i don't see any requests from haos in adguard home
dial tcp: lookup ghcr.io: no such host

Search this channel for ha dns and you should get some guidance
It's not DNS.
It can't be DNS.
It was DNS.
This is how dns issue looks like? 😄
core 2023.6.3 didn't starts, also like esphome.
however I was able to update the OS to 10.5 which come out this weekend.
What could be wrong on my side?
Totally possible dns
i see. how can I solve it?
Disable dns fallback and try setting another dns server
Hi,can i use a 2.5 inch ssd with my raspberry?
OK thanks
Huh. TIL that the containerized version of HA is missing a lot of important functionality.
Not really, no
Add-ons are hardly important functionality - pretty much anything there's an add-on for there's a container (or other way of running) for
i get the first part it is a line in the cli,
But a I don not the second. How do I set up an other dns server?
https://imgur.com/R8nOeIA
this is my dns info.
i am using a static ip
and I have a wireguard set up also, thats how a I reach the server from outside of the network
what should I change it to?
you could try 8.8.8.8 or 9.9.9.9
dns fallback is to cloudlfare but I See you have that disabled
yes I disabled it that was your recommendation. But it was enabled and didnt helped
where should I find dns on this page?
https://imgur.com/VP96gJa
how does it looks like?
I Would remove the 192.168.x.x entries
as clearly they aren't working for you right now
so remove them to test
Well, I am not close to the HA server right now, and I am afraid that I mess up the wireguard.
well, it's the next step
So, I didn't even know addons existed until today, because everything related to them is missing from the container UI
It's true that in principle I could set up addons myself, but that is a lot of work and research for each one, since the HA documentation assumes you're using HAOS and whatever daemons the addon is based on don't care about HA in the first place (or, to the extent they do, also assume you're using HAOS and not configuring everything by hand)
So at that point, it's probably easier long-term to figure out how to boot HAOS in a vm on my server and use that, vs. hand-configuring every single addon I might want
(Although I'm not very happy having HA running in what is effectively a black box, and backing it up is going to suck, but it seems that it's very much not designed to just run as a self-contained daemon on a larger system)
looks like it didn't helped...same error....
Well, you get to pick your tradeoffs...
But I don't want to remove 192.168.88.1, 192.168.8.1 I am affraid that I will loose the connection to the HA server.
I will investigate further if I am closer to the server.
changed dns to 1.1.1.1
ha dns options --servers dns://1.1.1.1
ha dns restart
Still the same thing, reboot doesn't help
well some users isp's block DNS that's not to their servers, so try your ISP's DNS and did you try ha dns options --fallback=false ?
because when HA can't reach the configured DNS it tries 1.1.1.1
I'm not very happy having HA running in what is effectively a black box
backing it up is going to suck
Contrary for me, I have made HA as turnkey as possible to reduce my "operations" efforts/time; HAOS on an SBC works very well for this (I'm using ODROID); NUCs too. Backing it up is very simple and automated with https://github.com/sabeechen/hassio-google-drive-backup.
If I had a spare sbc with more beef than a raspi I'd run it on that, yeah
But here we've got one central server + a bunch of rpi2/3s
Hey, I'm trying to install HA OS on a NUC, on an SSD that previously had Proxmox installed on. I want to install HA OS directly without Proxmox.
When I try to do it with Balena Etcher through Ubuntu (booted from USB), it fails, says "target ebusy resource busy or locked". Also when I try to wipe the partition through Ubuntu utility disk, I get an error.
How can this be solved?
HAOS in a VM will still minimize your efforts and intrinsically have the things you miss (about the same reasons I use HAOS). VM vs baremetal is same pros/cons for HA as w/other apps, so you should have success as sounds like you understand VM configs w/apps (e.g. accessing devices).
Yeah, I think the main thing giving me pause here is that I've never used libvirtd management tools -- qemu for one-off things, virtualbox for local stuff on occasion years back, but that's it. So there is a certain degree of straightforward appeal in burning haos to a USB or SD card, slapping it into an SBC, powering it up, and waiting for it to respond to ssh. Especially since it looks like I'm going to have to interact with virtd rather than just telling nixos what I want and letting it happen like I can with docker. Oh well.
None of this is urgent, so I may see if I can scavenge up a pi4 or a cheap off-brand SBC first before digging into virtd administration.
Yooo. Please help
So I'm so tired of virtualbox and my USBs getting disconnected once a day at random times.
If I move to Proxmox, am I basically starting from scratch with all my automations and devices?
still the same error and i don't see it reaching for 1.1.1.1
https://i.imgur.com/8qDHGoq.png
what does net info say
that says your DNS server is 172.16.0.10
then what ha dns options --servers dns://1.1.1.1 does?
Oh daddy!
If true, that makes me very happy
that is an OLD command
from like 3 years ago
lol
why is it a thing
what old information?
why people tell me to use it
because people don't know
this is why I don't like youtube guides
when they are outdated and people don't research to see if they are still valid 😛
today
even "Mod" lmao
he didn't tell you to use that command tho
he told you what to search for to find the answer 😛
search resoults would show only that command so yes he did
that's not true
Hello guys,
I am using Control4 with KNX in my home. But I've installed UPS and its parameters are being read via Modbus. Is there any way to integrate the Home Assistant in the Control4 for the visualisation? Is there any way?
Thanks!
yes that's true
FATA[0000] Error while executing rootCmd: unknown flag: --ipv4-nameserver
wtf https://i.imgur.com/qzGlElW.png
so no, he didn't tell you to use that command
there's 43 more results if you need me to point them out to you one by one
you can see here that fallback is false
ip4v != ipv4
thank you im blind lol
I've recently got a home assistant yellow. I just accepted and installed an update that popped up and after a restart its now showing the onboarding screen again. This happened once previously after restarting the yellow, but I didn't think much of it as I hadn't done much setup. I am able to restore from a backup, but I'd rather solve the problem if there is one.
there is a certain degree of straightforward appeal in burning haos to a USB or SD card, slapping it into an SBC, powering it up
yes, my "make it an appliance" desire for this was high (having spent enormous amounts of tinker time on other apps in the past... wanting to avoid again). dedicated box, no interference has helped.
waiting for it to respond to ssh
i almost never use console for HA, almost always UI only. Except for some issues diagnosis and solving, it's been nice being "more of a user" for once instead of feeling like "part of the product"! lol HA error handling still needs work, but overall it works decently.
Check your CM4. A module not completely seated in its sockets is a common culprit.
I didn't get a click when I was putting mine in. Am I supposed to? How hard am I meant to push it in?
It ususlly makes a satisfying click. Don't push with brute force, but it might need a bit to overcome the resistance
would getting an SSD make this now happen anymore? I've got one in the post right now
It should be good without SSD. If checking the CM4 doesn't help, head over to the channel #yellow-archived
got satisfying clicks this time on both sides hopefully its fixed now 💪 🙏
@tender wren Am I right that there's no way to recover the installation now (apart from a backup)?
I don't think so. I would suggest to reinstall the operating system, just to be safe.
I've got it up and running now off a backup. Is that technically reinstalling the OS? Everything seems to be working as it was before
It's not reinstalling the OS. But if it survives a reboot (not just a restart) without going back to onboarding, it should be fine.
no it's not quite the same as a reinstall
but I'd get those backups to a safe place JIC
I've already got the backup to google drive addon going 😎
that's what saved me just now as well
das ist gut.
Oh, I quite like the web UI. I just tend to use ssh for initial system setup -- set hostname, configure nfs mounts, etc.
(web UIs often don't expose that anyways, so)
(HA certainly doesn't, but maybe that's another thing that's excised from the non-HAOS version)
I'm having some DNS issues also.. but it seems like my issue is that the DNS container wont run. It crashes with an error log of:
panic: runtime error: invalid memory address or nil pointer dereference [signal SIGSEGV: segmentation violation code=0x1 addr=0x17 pc=0xc00057] goroutine 1 [running]: gopkg.in/DataDog/dd-trace-go.v1/ddtrace/tracer.(*span).setTagError(0x136bf00, 0x7, 0x0, 0x7) /go/pkg/mod/gopkg.in/!data!dog/dd-trace-go.v1@v1.28.0/ddtrace/tracer/span.go:131 +0x57 [11:16:17] WARNING: Halt DNS plug-in with exit code 2
I have tried pulling a new package from ghcr.io, but it hasnt had any effect
I did also try a thread on the topic but havent had much traction, https://community.home-assistant.io/t/default-dns-cant-resolve-anything/604859/5
Hi, I have a problem with USB Passthrough and proxmox. I bought a pre-configured dell wyse terminal with proxmox and home assistant virtualization installed and separately a zigbee module. The proxmox is standing on a static IP address of 192.168.1.25 and my router's range is 192.168.0.253. I'm wondering if anyone may have had experience with proxmox and changing the static IP address or if I will be able to set up USB Passthrough or change the IP address so that I can remotely change the settings after connecting a keyboard and monitor to the terminal. Alternatively does anyone have any ideas for another solution?
If you connect a keyboard and monitor to the device directly you should be able to change the proxmox settings there
You don't need to pass through anything for proxmox itself
how to change locals dns, because haos doesn't want to use configured servers nor use fallback?
Hi all. I'm looking to install Home Assistant using the Docker container. The installation guide seems to insist on host networking and privileged mode.
Are both of these mandatory? I'd prefer to run it on a bridge network with no exposed ports, with an Nginx reverse proxy managing the frontend access.
you will lose discovery features and maybe experience trouble with certain integrations but you can do it if you want
Why is that, I'm assuming because multicast will not work correctly?
yes
Okay, thanks. I'll see how it goes. I already have my ESPhome running in a container, for that I use an mDNS relay which seems to work well. I could reuse that for HA too if too much breaks.
That is a bad idea because these reflectors often times do more harm but you can try
Also if you ever plan to use Matter or Thread you'll run into problems
I set myself up with the DuckDNS add-on with a Let's Encrypt cert, and I can visit https://mydomain.duckdns.org:8123 from the browser and it works just fine
but when I plug that into the Android app I just get a black screen on the app, it doesn't even pop up an error anymore letting me change the server url
I can visit the site on the phone browser just fine
what am I missing here?
okay bad data apparently I am missing something... I can visit https://<box IP>:8123 in the browser but https://mydomain.duckdns.org:8123 gives me connection timed out
likewise on the PC, the duckdns url times out
so I spose it's off to the router to investigate that NAT loopback is enabled
alternately I could override internal DNS
that means you didn't set up duck dns properly
if you use duck dns, you should only have access to your domain name, not your IP
unless you set up nginx
Also, the duckdns addon manages your cert
So you don’t need the letsencrpt addon
this is what I do
here's my DuckDNS serttings, not sure what I'm missing
https://pasteboard.co/XnOrD69TTqaK.png
did you do port forwarding?
I spose I would, even with NAT loopback
pretty sure it's covered in the installation procedure
right because the webserver is set to respond to <IP> instead of <whatever.duckdns.org>
yep
well let's see if the android app is any happier
also, you could try a croudflare tunnel avoid avoid this all togther
but it has limitations, i.e. you'll be banned from croudflare if you stream video
solved
no worries, my home network is built business-style with Arista UTM and UniFi throughout
so I'm not really feature-limited like a home user router
I was just pointing out that it's against croudflare tunnel TOS to stream video
good point
so if you don't want to port forward, but you want access, you can use a croudflare tunnel
or you could pay for nabucasa too
if i wanted to do the tunnel thing with video streaming I'd just set up an exit endpoint on a DO droplet and do it myself 😄
Has anyone installed HA on a raspberry pi kubernetes cluster? Should be pretty straight forward, my largest concern however is my conbee II usb device....
A coordinator that's connected over the network, rather than USB
a zigbee coordinator?
Well... of course?
can I turn another rasp pi into one?
With skill, yes
haha i mean like something you can buy? or have to build
aww snap
so pie cluster aside..
are there advantages to these?
vs a usb stick?
ill move this convo
They're away from all the RF noise of your Pi
Dannnnnnggggg
Alright back to HA installed on a pi cluster... probably wont see any kind of performance ganins. Really just redundancy I can count on?
Yes, assuming you've got the file system failover handled
Dont know the exact wording.. but currently have had HA running on its own PI 4 for years.
Will it run worse or will I lose anything by running in a container vs being the only software on the device
I have a question. I'm trying to install homeassistant on my raspberry pi 3. But i got an error. WARNING (MainThread) [supervisor.jobs] 'updater.fetch_data' blocked from execution, no supervisor internet connection. Somebody knows what the problem is and how i can solve it? Im new to this so i have clue
are you able to get to the console?
bump
Not without googling how🙈
Enter network info and supervisor logs and show us the results
Dont know the exact wording but
Well i have connected a keyboard to it but cannot type
I have connected it to another ethernet cable and now i doesn't give me an error on the supervisor
So i think i have to wait now for 20 minutes or so
Well it works now. I will just keep it on this cable
Thanks for helping
sorry got distracted
Sometime this week will be migrating from HAOS to a container version of HA.
Will HAOS backups help at all? Cause issues?
You can manually extract the contents for HA (Core) from the backup
So that will be some sort of option when I go to upload the backup?
or something i need to do before I expose the backup file to HA (Core)
manually extract the contents
The backup is a zip
Unzip it and you'll get a bunch of .tar.gz
One is for HA
Extract that to your chosen location
And then set up backups
So you mean like manually backup any .config files
Well, by manually you mean automatically outside of HA
ok
And your Docker config
then just copy/paste when im starting over with HA core
You have no idea...
I will have to rebuild my zigbee devices
And you set up backups of your config folder to get all of that, and more
Alright im able to download my entire config folder now
once booting up HA core, I will be able to restore from that? Or most things?
Same as when you asked up there... no
Only HAOS has restores
You have to do this manually
Can I suggest you reconsider this until you know what you're doing?
Set up a test host with Docker, learn how to back up and restore
Practice ...
Leeroyyyyyyyyyyyyyy
Learn how to do DR
Saved an entire copy of:
addons
backup
config
media
share
ssl
that should cover me
I also just run to enstences of HA till if get everthing swapped over
Hi, anyone experiance with running Fedora Linux 38 with KVM and trying to get a HaOS VM running? I cant seem to get the networking running.
I using Generic Linux as OS for the VM
i meant, like ubuntu
Fedora Linux 38 is the OS KVM is ruuning on
or something other than ha os
Yes, i did. Then it just doesnt boot
just to clarify, the VM does not boot?
It does boot. But the networking part is not working. https://i.imgur.com/XxEmxBe.png
yes, and i was suggesting to try ubuntu or debian as the VM
to make sure it is not home assistant that is casusing the problem
and did the VM have a network with a different os?
if not then you might want to check your host. Maybe the KVM bridge is not setup correctly.
It was rmmod br_netfilter for the bridge settings i guess
Weird af
Never seen that before..
any specific reason for fedora?
if not then there are other distros out there like proxmox which have a reliable virtualization platform
fromthe screenshot i can see that the VM does have a interface assigned to it. I can only guess that the kernel module for the linux bridge is not configured properly.
it could be that the bridge does not have your primary interface as one of it's members
in which case there would be no IP address.
Would I be wrong for thinking Home Assistant Core should match what is shows for supervisor version?
https://imgur.com/sVtXAhO
These are different things: https://github.com/home-assistant/supervisor/releases
Core version is in the bold line
Hi, I would like to install HA supervised on an old atlon based mini-laptop for a proof of concept. I am following this guide: https://github.com/home-assistant/supervised-installer. I Installed Debian 12 from netinstaller over ethernet (not WiFi). All goes well up to the point where I have to manually install udisk2 in step 1. The other programs install fine but udisk2 says (translated from Dutch): 'Cannot find package udisk2'. I have very limited knowledge of linux but I can follow orders 😉 Any hints are welkome!
udisks2 not udisk2, by the looks
Omg, you are right, stupid mistake. Thank you so much!
Assumed it was just a typo 🙂 Easy one to make
Unfortunately I still get lots of errors on several files: Temperory failure while searching 'deb.debian.org'. (again, translated from Dutch, mamybe the english phrase would be slightly different)
Apt-get update is suggested, but doesn't help a bit, and 'apt-get --fix-missing'' shows me the same errors.
Ok, networking top to bottom:
ip ato see if you have an IP addressip rto check for a default route (default via x.y.z.wline)cat /etc/resolv.confto make sure you have at least one DNS server (nameserver x.y.z.w)
ip a show among others: enp1so .... state up and my internal ip for thais machine. ip r also shows a bunch of stuff like the machine ip again and 192.254.0.0/16 scope link metric 1000
and 192.168.1.0/24 dev enp1s0 proto kernal scope link src mu-ip
cat /etc/resolv.conf shows : nameserver 127.0.0.53 ; options edns0 trust-ad ; search .
nothing more.
I see I can ping 8.8.8.8 but I cannot ping google.nl. So I guess there is a dns-problem?
Oh good, it's my old friend systemd-resolved. Try replacing what's in resolv.conf just with nameserver 8.8.8.8 and see if you get any further
I did: nano /etc/resolv.conf. Th file holds my gateway IP. I replaced it with 8.8.8 as you said. Should I reboot? Tight now it doesn't help...
Reboot didn't help either. I find it strange that I have a dns-problem while the other packages were installing fine!
And I can browse to the requested files with my other computer too, so they are on that server...
I'm rapidly running out of easy guesses and quick hacks, sorry. It's also midnight where I am so afraid I'm also going to have to try and sleep at some point
there is a octet missing in your ip
thanks for helping, have a nice sleep!
Yea, but that is only here in discord, I had them all in the resolv file. (just checked again but after the reboot, my gateway-ip is in again)
changed it back again, checked the four numbers, same result.
Hello everyone yesterday i tried home assistant supervised install on an old netbook with debian 12. I folowed the official installation on github and I install it successfully. When I go to the url it is showing preparing home assistant and when I click to view the log I have this error : 23-08-23 16:29:49 ERROR (MainThread)
[supervisor.docker.interface] Can't install
ghcr.io/home-assistant/qemux86-64-
homeassistant:2023.8.3: 500 Server Error for http+docker://localhost/v1.43/images/create?
tag=2023.8.3&fromImage-ghcr. i0%2Fhome-homeassistant&platform-linux%2Famd64: Internal Server
Error ("Get "https://ghcr.io/v2/": dial tcp
140.82.121.33:443: connect: no route to host")
23-08-23 16:29:49 WARNING (MainThread)
[supervisor.homeassistant.corel Error on Home Assistant installation. Retry in 30sec. Plz help me I tried to solve this 2 hours I am a beginner on Linux
is there a reason you went with the most difficult install method ?
probably because it's an old notebook
Yes I want to use plugins and with hope assistant core it is note possible
No it use amd64 architecture
then use HAOS
well that I understand, but also OP said they are a beginner at linux
Don't go supervised, espeically if you don't know linux
I can t because my pc doesn’t have ufei
this is what I Was getting at
ah ok
Is there a solution to convert ufei iso to legacy bios
not sure, outside my expertise
you might need to make a forum post about this
although a wiz is about to respond
or not?
or yes, he's typing again
you can run a VM without uefi
No it,s not me the my cpu is the intel atom n150
oh atom
I thought I saw athlon when I was on my phone earlier and my first thought was that I hadn't even seen that word in a long time.
amd64 you saw
Yes but it,s same as x86_64
you don't have anything laying around from 5 years ago instead? 🙂
the CPU in my box is from 07 cogneato 😛
If you're having probs getting linux on it there will be more problems down the line.
No I only have this
#templates-archived can help
oh sorry, wrong chat... 😫
Hi there! I´ve been stuggeling with HA peaking on 100% CPU. I run HA OS in a VM on a Synology NAS (.ova), and I havedone a fresh install, putting back my backup twice with seemingly no effect. I don´t find what´s causing this, and started to disable Integations. No luck so far.
Had it installed before, thin I´ve removed all of it...
is it a j series synology?
220+
yeah our j series is a joke
Have allocated 2 CPU´s and 2 Gb RAM for the VM
the 220+ isn't much better, it's got a 2-core Celeron J4025 at 2.9GHz and 2GB system RAM
probably not enough brainz for the project @sage kestrel
that would be the entirety of the computing resources on that box, dude
oh okay, well you're still giving it ALL the processor
all that beefy Celeron goodness we've come to expect from 2019
synology changed their OS a while back too which make it more power hungry
at least for my system
IIRC I've got a 216J?
I have not been impressed by Synology lately
so I should run the VM on one CPU, then?
give it a shot, that'll at least free up enough brainz for it to be a NAS as well
donno if it's gonna be enough grunt for home assistant but it's worth a shot
Bu I thought I somehow allocated some of the CPU only, see now that was wrong...? Or are there fragments of CPU in there?
ours is the ds220j and it's terrible
that celeron has 2 cores
minw has 2x 2GHz CPU
and while you can virtualize 4-6 cores on a 2-core machine, it really ramps down the computing speed
INTEL Celeron J4025
and 6144 MB RAM
How woul I (accidentally) have done that?
... how can I put it back?
just assign the home assistant 1 core
if you've assigned it 2 cores then the machine is busy using its two cores to run all the NAS and Synology stuff plus virtualizing two more cores
and it's going to struggle at that with high CPU utilization like you said in your initial post
OK I give that a shot. Shutting down VM now to change it
Than´s for help so far 🙂
I know it could be several reasons why HA peaks on CPU, as Integrations, and processes that hang.... That would seem more reasonable though 🙂
yeah in your case my suspicion is that you just don't have enough compute resources to give it
if it works better on one core then it should do you fine for a while
Hope so too
Th strange thing is that on th average CPU for HA is seldom more than 50%
yeah but think of it this way
you have two physical cores
you're asking those two cores to do all the NAS and Synology stuff
and then you virtualize two more cores on top of it
so now they're doing the work of four cores
of course the CPU utilization will be high
now, you've got home assistant on one core, all the synology stuff on the other core
the CPU load should be lower
This is how it was today: https://paste.pics/d9d49ab0921bb412143aecbdeca9ea71
at the moment, setting up Docker Application Container Engine it´s 4%
a task tha´s been going on for 17 minuites now
And I also got the CPU_IOWAIT, which f course would be due to lacking resources as you say.
seems normal so far
Good
fairly high utilization for you is going to be a fact of life, there just really isn't that much grunt in the Celeron processors
21 minutes setting up DACE now...-
and looking at the spec sheet, I suspect you're sharing some of that 2GB system RAM with the graphics
how?
from the Celeron processor spec sheet
Graphics Video Max Memory
The maximum amount of memory accessible to processor graphics. Processor graphics operates on the same physical memory as the CPU (subject to OS, driver, and other system limitations).
This was earlier today: https://paste.pics/81ffb62dfd9786647eb9560b638d8ff2
yeah
if you're still getting them, that tells you the synology just doesn't have enough grunt for the task and you should find an old dell or something
I also get the Emergency CLI loaded at startup
404 Not Found
does your mDNS not work?
irrelevant
can you please stop tagging me on every reply
I don't use discord daily, it's default option for a reason
Hey everyone 🙂
yeah because discord devs are brain dead
Generally, don't tag people to ask for help - it comes across as bad manners, you’re demanding somebody answers you. It’s different if you’re thanking somebody, obviously. If you do tag somebody keep it polite and respectful. Remember that everybody is a volunteer, and nobody has to help you, and people may block you.
Similarly, please don’t DM (direct message) people asking for help. It also comes across as demanding, and means that others can’t learn from what you do.
Finally, please keep tagging people in replies to a minimum. That too can become annoying very quickly and should be used only when it's necessary (such as if it's been a long time, or there's multiple conversations going on). When using Discord's new Reply feature it defaults to pinging the person you reply to, click @ ON to @ OFF to stop this - on the right side of the compose bar.
I am running docker compose.
Is updating not just
docker stop homeassistant
docker rm homeassistant
docker compose up -d
?
Docs are just for normal container, not for docker compose
No, you're wrong
there
Look at the link that was posted
Also, the main Docker docs also cover Compose for the install
Ah, sorry, it is not explained in the actual RPi area how to update compose. Only normal docker. It is "hidden" in the general docs
well its not that well hidden if I found it in 0.3secs
Well, as RPi user you tend to check RPi instructions because RPi does have quite a few limitations
over pigeon internet
can i buy a pixel pat?
Yep, but you would not have found it as an RPi user looking in "the obvious docs" 😉
no but I found it in the docs that were linked....
Well, yes but no
To update you have to restart (well, replace) the container
You can (and have to) pull first
But no need to remove the container before pulling, I assume?
Btw, the docs state docker-composebut it should be docker composewithout the hyphen
Feel free to suggest an edit
that's also relatively new
Oh, I still have docker-compose.yaml everywhere
So do I
"docker" is being expunged everywhere in favor of "container"
it used to be practically synonymous
Most recently, Synology renamed their "Docker" addon to "Container"
How long is HA restore from backup meant to take? Minutes? Hours?
(It's a backup from the docker version, being restored into HAOS, and is about 8MB total size)
Oh, hmm, looks like the web ui crashed.
to what hardware?
HAOS running in qemu.
on what hardware tho
Host is Ryzen with 64GB of memory, forget what motherboard. Guest is 4GB memory, 2vCPU.
should've been really fast with such a small backup file and that hardware
maybe it's networking related like the static ip + ip reservation not being released?
Restores always seem to be slow these days
Dunno, just that there's been more posts of late about slow restores than are usual
More posts about problems with DNS too
Probably gremlins
DNS is a weird and crazy beast anyway
haiku time
It's not DNS.
There's no way it's DNS.
It was DNS.
DNS is working.
It must be the firewall rules.
Let's open them all.
Ok, I can figure out a bit more of what's going on.
I tell it I want to restore from backup during onboarding and offer it the tar, it lists it as a "partial backup" containing data from homeassistant core, I tick the box next to that and tell it to proceed.
The web UI immediately stops responding.
The VM is still up and I can log in to it.
Looking at journalctl, it failed to restore the whole backup, because it ended up with some named pipes rolled into it.
However, it clearly restored part of it because I can see log messages from it trying to access cameras and stuff defined in that backup (and failing, because it's confused about its hostname and dns search path)
Let's try this again with the pipes pruned out of the backup.
Aha! Ok, there were two different issues going on -- the pipes, and the fact that the restored configuration.yaml had an http block that wasn't compatible with the vm.
This haiku, I like it
Without DNS
We would have to memorize
Numbers for websites
DNS is a map,
that guides us through the web's maze,
to our destination.
One of my teammates at work was being subjected to a mildly-incorrect but very 101-level explanation of DNS from someone a few days back.
I got to bust out this piece of internet history (which I think is possibly older than they are, and yet still so relevant): https://archive.ph/DlhDg
(not entirely SFW, just so that's mentioned before anyone clicks on it)
This particular person is young enough to have never used a VHS tape, floppy disk, or dial-up internet. Boy did I feel old when we found that out
I took immense delight in explaining CGI scripts to them though. The look of horror on the face of someone when you explain "so the web server just executes this file and sends its output back to the browser" 😄
Speaking of DNS, something appears to be badly wrong with HOAS DNS for me
ha dns info shows locals: dns://192.168.1.1, which is correct, but also host: 172.30.32.3, and looking at the logs it appears to be sending all of its DNS requests to the latter, which obviously does not work
I tried turning off fallback (ha dns opt --fallback=0) based on some forum posts, but it does not appear to have had any effect even after restarting.
Also, looking at the logs, it's trying to look up hostname.local.hass.io rather than using my local DNS search path, even though its hostname shows up correctly as hass.mydomain.ca.
any open issues?
As in reported by ha resolution info? Only "no current backup", which makes sense since I just set this up tonight.
no as in on github
https://github.com/home-assistant/operating-system/issues/2699 people have been having haos dns issues lately it seems
i'm on haos without any dns issues myself however lol
That's the opposite problem of what I'm having
that's cuz it's the wrong command also, it's ha dns options --fallback=false not =0
it accepts 0 as a synonym for false and now shows fallback: false in ha dns info
That's the exact opposite of the issue I described here. Their problem is that it gets DNS info from DHCP and refuses to use anything else. My problem is that it ignores DNS info from DHCP and is using some random DNS server on the wider internet that can't resolve hostnames internal to my LAN.
i just linked to an open dns issue for haos
i wasnt specifically stating that was your issue
rather suggesting you make sure there isnt a known issue before going down the rabbit hole
Yeah, I've rummaged around in the github issues and found nothing.
Interestingly, nslookup and resolvectl query (on the hoas box) both work as expected.
So the problem is specific to HA, not with the system's network configuration as a whole.
i'd just start a new issue. the haos devs dont really hang out in this channel (more in #beta or dev related channels, or just on github)
agners and others will jump on it - and the more technical info you provide the better
(this also implies that HA is not using the system resolver???)
docker exec -it hassio_dns bash
for fun you could take a backup then try the 11.0 beta
try that from cmd line
and see if they've made some changes
doing that and then running nslookup from inside the dns container also works, and shows it using 127.0.0.11 as the nameserver (which is presumably the dns relay running inside that container)
Which makes me suspect that the root cause is actually that it's blindly appending .local.hass.io to every hostname it tries to resolve, and when the local DNS can't resolve that it falls back to someone else's DNS rather than trusting the local nameserver to recurse
ha dns reset
ha dns options --servers dns://8.8.8.8 --servers dns://1.1.1.1
ha dns restart
been thru this?
The search path in /etc/resolv.conf is correct, though, so where is this happening?
I haven't, and also, how would telling it to use Google's DNS help at all?
well I'd try your ISP ones first
