#yellow-archived
1 messages · Page 12 of 1
I'd still ask. You can still consider to do it this way
Hi all!
I have a pre-installed home assistant yellow, configered literally everything, and decided: lets move it to its final destination.
unplugged it, plugged it back in, and now it is in a strange state. I cannot find this blinking in the manual: the yellow led is constantly blinking twice, the green is constantly blinking once (no 2 second gap as in the manual). See video. What could be wrong?
I can ping the ip-address and going to the home assistant url gives an connection refused.
I do have backups of al my setup, but these are on the device, so I am reluctant to do a factory reset since I spent hours configuring everything 😅
@mighty oar you can connect using the serial console option to check logs https://yellow.home-assistant.io/documentation/
Under troubleshooting
Is the final destination a completely different network or just a different physical location on the same network
So my yellow arrived and it would never boot and i couldnt access it. Open port scans showed no ports open.
I used the rpiboot method here to reinstall. https://yellow.home-assistant.io/guides/reinstall-os/#install-os-rpiboot
I was able to get the yellow installer on the eMMC, but now I see it on the network from firewall/dhcp, but can't ping it or access it or anything. My red LED is on, green flashes every once in a while. Trying to access it via http://10.0.0.144:8123
https://i.ddy.moe/2023/11/vivaldi_231115-075437.png
notably, the hostname is still yellow-installer, not homeassistant.
Aaaand now it's not actively connecting to the network. :|
Can I just directly flash it with HA OS instead of the yellow installer?
@eager mortar yes, and that's what I prefer to do myself. I have had issues flashing the installer to emmc
But never when flashing the image for yellow
But I'd check if the CM4 is fully seated
Is there a yellow specific image?
The yellow xz file I got from GitHub put the installer on it
haos_yellow-11.1.img.xz
https://github.com/home-assistant/operating-system/releases
hmm, maybe your flash didn't take then? It is not the installer
And yes, check that the CM4 is fully seated. It can be deceiving
i assume it's seated cause i can flash it and such
but i'll check after retrying this
okay yeah this is going much more slowly than before
maybe i still had the installer image selected
lol
from before
whoa
You can flash it, set up HA, and then still have a partially seated CM4. An indicator would be that after a reboot of the yellow itself, you are back to a fresh onboarding page as though you never set it up. But that's one of those wait and see if it happens issues
is there a nice way to remove the heatsink?
You're done with the flashing?
you can press down lightly on it, causing the pin to push further out, and then squeeze the "arrows" together with your fingers or tweezers a bit, then push it back up through. You only need to do one pin and then it can be rotated out of the way, but I'd save that for after
i'll go thru that process if it gives me trouble, i think
no reason to stress plastic parts if it's not an issue
I agree. And it can pull the thermal pads off
I would go through onboarding, skip restoring at setup, get to the dashboard, then try a reboot of the host
If you get back to the sign in page then you should be fine
Option 2 here, step 9, should just say to install the os directly, not "flash the Home Assistant OS Installer for Yellow". i feel like if you're going rpiboot at that point, it's easier to just go all the way.
same for here
I agree with that too, and don't understand why one would bother flashing the installer, but need to confirm the reason for it with @shut cedar
The installer also runs a firmware update afaik
reboot successful 👍
Are you going to restore a backup, @eager mortar ?
different location on the same network
ah right, yes
If you create an uncompressed backup on your old machine (with a service call hassio.backup_full in the development tools/services - and deselected compressed backup) you might have a bigger backup file but save time during the restore process
hm okay
is there any reason i'd want to exclude the HA databse from the backup?
lol this is funny
Without database the backup is smaller, but the history is gone
makes sense
okay last question, how do i download the full backup off the old device?
i only see a delete option
ah
clicking it
Main reason is we would need to add the OS image itself to the RPi imager, which I am sure people would mix up (install HAOS onto the USB thumb drive) 🙈
In the end that installation method is the exceptional case anyways...
think it would still make sense to document it
even if it's not the default
oh in fact you have a troubleshooting page for it. just linking this to the other page would be a good option. https://yellow.home-assistant.io/guides/issue-while-installing-os-on-emmc/
Yeah I am fine adding it to the docs of the rpiboot case 👍
When I've tried flashing the installer directly to emmc it has failed both times.
Hm I see, also didn't remember that 🙈
any idea how long i should expect a 1GB full backup restore to take?
Long enough to find something else to do with the next 30 min. It could be done faster or slower, but I'd do breakfast, lunch, laundry...something
And don't panic if it takes longer - just let it do its thing and try to call the HA UI from another browser tab once in a while.
If you enable advanced mode for the user, then go to the logs page and choose supervisor logs from the dropdown in the upper right, you can see some progression
Looks like I'm restored. Have a bunch of things I need to reconnect tho, I think
Oh boy, trying to set up zigbee has cashed my HAY lol
@eager mortar it's good to full reboot after a restore to get everything starting as it usually would
yeah well didnt have much choice there lmao
not a restart of HA
but the whole system. It doesn't do that automatically after a restore
(and add-ons often behave weird after a restore without reboot)
I had to power it off to recover
ah
Hi is there any mechanism to monitor the health of the nvme drive attached to the yellow?
You can use the USB-C serial console or the debug SSH to access the OS itself and use the nvme command
I'm not trying to debug any issues. Rather asking if HA has any built in health check mechanism or not
Currently no, NVMe health data is not being read out.
Considering they are used as a data disk, it's going to have high number of read and writes. Might be a good idea.
You can submit a feature request here:
https://community.home-assistant.io/c/feature-requests/13
Has anyone tried to use HAOS for Yellow on a SK Hynix Platinum P41?
At least the SK Hynix BC711 is incompatible
i managed to brick my yellow and have now tried to reinstall it. Everything seems to worked out ok but i can't access it through the network. Using the console via usb i see that the network does not start or at least does not get an IP., any tips?
How did you brick it?
From console https://pastebin.com/acQPTeWV
Jorg: not sure exactly what why it happened but it got stuck in boot mode. started up with only red power led.
however seems i finally managed to get it working again, had to swap ports on the router
but it is still not responding on ip:8123 or homeassistant:8123
So it has an IP now?
seems like it just needed some time, the web page loaded now
after this experience i wonder why the documentation almost expect me to be using windows or a mac to recover. It really needs a linux section as well 🙂
I only use bricked if it can't be interacted with in any way. Like, dead dead
What exactly in the docs are you referring to?
Jorg: the part on rpiboot, was not really working out of the box on fedora at least
am i the only one with heat problems on my yellow ? it freezes up because the cpu and nvme gets to hot
If the CPU gets too hot, the CM4 shouldn't freeze but throttle down - so that's not a normal behavior.
maybe it throttles down but it takes minutes to load pages so i say freezes
Are you running it in a hot environment or does it get hot by itself?
works fine for the first hour but after installing everything it get worse and worse
nope ambient between 10 in the winter to 22 in the summer maybe 25 at max
it gets hot by it self i would say
Mine (the Yellow) is running at 45.8°C now and didn't peak above 60°C in summer (with high CPU load).
Did you get a Yellow version with a preinstalled CM4?
If checking the CM4 and the NVMe for proper seating in their sockets doesn't help, I would suggest to contact support at yellow@home-assistant.io .
Describe your problem and let them know how they can contact you here.
no i installed it 2 month ago when i got the cm4 almost 2 years of waiting 🙂
ill TS some more
Does anyone have suggestions for a safe place to buy a power supply? I'm nervous buying electrical components on Amazon these days.
hello all i try to migrating to HA yellow form a RPi 4 but i run in to a problem so i hopw some one can help sorry if the writing is not all correct i am form holland
me HA is running on the yellow and the add ons too but in esphome i cant see annything all the intergrations are installd i see all the entiteiten in node red i can use the entiteiten but i cant edith the yaml file's
if i connect a new divice a d1 mini to the usb of the HA yellow i can start the add new divice but can not sellect d1 mini only esp 8266 it start te progress of installing gets to the end is wil see the wifi network but afther that it runs in to a fail
[17:16:53][W][api.connection:091]: : Socket operation failed: BAD_INDICATOR errno=11
Have you rebooted the Yellow (the complete system, not just a HA restart) after restoring the backup?
Yep i did that and did a new install too get the same problem
and if i restore a backup then esphome will not start up so i try from a fresh start
Is just ESPHome causing trouble? What YAML files cannot be edited and what are you using to edit them?
yep only ESPHome is causing the trouble if i run the back up all automations work but i cant get ESPHome to run so cant get to the YAML file
So you mean the ESPHome config files, which are edited within the ESPHome add-on
In that case I would create a copy of the directory /config/esphome , uninstall the ESPHome add-on and re-install it afterwards.
If the files in the directory have been removed, you can restore them from the copy you made before.
yep did that all ready and got the same problem
What do the logs of the add-on and the integration say?
at the moment i juist got a fresh instalation so the log is not saying anny thing beside the things i did
No error messages in the log section of the add-on or in the HA core logs?
only the frimware is not availebl from all d1 mini bord that i got
That's not a Yellow issue.
Maybe the people in the #diy-archived channel or at the ESPHome Discord might be able to help.
the thing is if i conect my RPi 4 all is working and i can see all in ESPHome
Without any errors in the logs and everything else working, I see nothing indicating the Yellow to be broken.
By the way - The D1 mini can be selected via ESP8266 - WeMos D1 R2 and mini in my ESPHome instance
The Generic config has to be changed using the dropdown menu
first off all verry thanx for try helping out
if i select esp8266 i wont get a dropdown menu to select WeMos D1 so that is strange to
Ah... you have to remove the checkmark at "Use recommended settings" when choosing the device type.
ahhhhaa now i got the drop down
lets see what happens if this runs then i can rebuild everything
nope to bad same error
[D][api:102]: Accepted 192.168.188.51
[18:25:58][W][api.connection:091]: : Socket operation failed: BAD_INDICATOR errno=11
As I said, the people in #diy-archived might be able to help you - there are more using ESPHome.
oke thanx again for the help so far i go try there
Tried to factory reset Yellow. It seems stuck, with the yellow blinking light for the last two hours. It can't take that long?
Options?
You can use Option 2 of this guide. But instead of flashing the HA Yellow Installer, I would use the HAOS image for Yellow (not the Pi4 one) from GitHub and an other flashing software (like Balena Etcher or dd).
https://yellow.home-assistant.io/guides/reinstall-os/
I went with the Yellow to avoid all these types of steps. Very frustrating to begin this process.
Thanks for the answer but I'm not comfortable enough to dive deeper into the unknown (for me)
You can also have a look at the USB-C serial console for error messages.
Hello, I'm a new HA user and have a POE Yellow. I have the OS installed, however I cannot get the Zigbee radio to work with ZHA, or Z2M. Initially ZHA would just give me an error in the UI anytime I tried to configure the service. I attempted to install the latest silabs firmware after googling a bit (NabuCasa_Yellow_EZSP_v6.10.3.0_PA32_ncp-uart-hw_115200.gbl), which went to 100% and then seemed to get stuck. I sat there for many minutes and then I rebooted the device. After reboot, same behavior, no integrations working. I tried to get Z2M working. I installed mosquitto, and Z2M, but Z2M fails to start with a 502 error. Around this time I realize it may be the fact that I had also enabled the Matter controller and did not have the multi protocol add-on. I added that and disabled Matter, and after some time of sitting, I tried ZHA again and it did configure this time! Great, (I thought) but now it shows "failed to set up" on the configuration page, and I still cannot start Z2M even when I disable ZHA. So I'm thinking there is a hardware problem? Also noteworthy, the blue button on my Yellow was broken out of the box and is non functional.
Can someone tell me how to look at some logs that may help identify the problem? I'm not sure where to look in HA
Are you sure you installed the correct version of the OS? Specifically the HA Yellow OS? I had a similar issue when I first started with my Yellow because I accidentally installed the RPi version of the OS, instead of the Yellow version.
If you look at Settings - System - Hardware, does it say Yellow or Raspberry Pi 4?
I just looked back in history and it was you Jorg that helped me solve this issue a few months back. Thank you and thank you for continuing to help others here. My HA yellow has been 👌
It does say Yellow
not sure if I should see anything different, but when I click on 'all hardware' and click on ttyAMA1 I don't see much. Not sure if I should see more than this as ttyAMA1 is always referenced as the port
DEVNAME: /dev/ttyAMA1
DEVPATH: /devices/platform/soc/fe201800.serial/tty/ttyAMA1
MAJOR: '204'
MINOR: '65'
SUBSYSTEM: tty
TAGS: ':systemd:'
USEC_INITIALIZED: '6089299'
The OS is correct. If ttyAMA1 does not show up, it sounds like a not fully seated CM4. Meaning it might not be completely in its sockets
It does show ttyAMA1 on the list
just not sure if its populating all info that it should?
That looks fine. I would start to remove all Zigbee/Thread related stuff. All integrations and add-ons.
After that I would install the Silicon Labs Flasher add-on. Select /dev/ttyAMA1 as device from the list on the add-on's settings page, save it and run the add-on. Look at logs page for errors. If it stops without error (stop code 0 is a good sign) remove the add-on and reboot the Yellow
Couldn’t get the flasher add on to start after selecting AMA1. I’m resetting my yellow back to factory and will only install the flasher first and try that
The flashing tool worked after a fresh reset of my HA! Going to try setting up z2m again
Look if there is a ZHA integration active in your integrations overview. If so, remove it first
After lots of failure, I was able to successfully configure ZHA and add a zigbee device! I started with ZHA to try and eliminate as much complexity as possible to just POC that things work
New question. Can I / how would I go about enabling Matter, while also retaining ZHA/ Zigbee devices
I just added the Matter integration, and then added a device via scanning QR code. It seemed to start the process fine, the device itself indicated it was starting to pair (LED light stopped blinking red and went to steady white) but the app gives an error stating the device couldnt be added to the network, and I should reset the accessory and try again.
What type of Matter device is this? Thread or WiFi based?
You can read about Multiprotocol support and how to enable it at https://yellow.home-assistant.io/documentation/. However, currently we have quite some reports that it is not really stable.
I just got the Yellow a while back and am totally new to smarthome things. What's everyone's favorite place to find example setups and show-and-tell for the yellow?
@fervent linden there's not much show and tell for the Yellow itself, but rather Home Assistant in general. The home assistant running on Yellow is the same home assistant running on every device, really.
The community forum is a good place to explore
Thanks cogneato :) I'll check out the HA forum!
Hello! Got the Silicon Labs Multiprotocol add-on, and wonder, where do I find its socket:// address? And how do I put that socket:// address into zigbee2mqtt?
- Silicon Labs Multiprotocol 2.3.2
Alternatively, is there a coherent zigbee2mqtt guide anywhere?
Or can someone explain the following here:
https://github.com/home-assistant/addons/blob/master/silabs-multiprotocol/DOCS.md#zigbee
"Remember/copy the hostname of the add-on (e.g. c8f00288-silabs-multiprotocol)."
Where do I get the hostname of the Addon? And how can an Addon even have a hostname? It's terribly confusing.
It's on the addon info page
And it's a container
And there's a DNS container managing the names
Found it, now it makes sense 🙂 Thank you
Zigbee2MQTT finally started with the 'Zigbee and OpenThread multiprotocol add-on' enabled. Wish they had just listed these magic settings instead. Anyway here it is:
mqtt:
mqtt: mqtt://core-mosquitto:1883
serial:
adapter: ezsp
port: tcp://core-silabs-multiprotocol:9999
Half a day later, with no HA experience whatsoever, got Zigbee2MQTT working through the 'Zigbee and OpenThread multiprotocol add-on'. Paired 6x Philips Hue lights, 2x IKEA Tradfri lights, and 2x IKEA Tradfri ON/OFF buttons via Touchlink. Pretty cool that the Tradfri On/Off works with lights, while the lights are paired with HA.
Doing homeassistant for the first time. Setting up a brand new Yellow (PoE kit) with a CM4108000 (lite, no emmc), a 30-watt trendnet PoE injector, and a sabrent rocket NVMe SSD. I did the usb-stick imager, the yellow light dance did everything it should, and reached the solid red power-down state. I unplugged ethernet, then unplugged USB. Re-plugged ethernet so it would boot. I get the setup webpage, but it just gets stuck and hits an error eventually (but the logs show no error). I have connected serial over USB and watch it seem to boot up into login prompt without any errors.. only to start the reboot procedure (stopping services, unmount, reaching reboot.target) about 6 seconds after reaching the login prompt. I've yet to get to any setup page. For that brief 6-second period, I can browse to it's IP:8123, so it is connecting to the network. Absent any errors.. I have no idea what to do differently here
its just a permanent boot loop, without me ever successfully performing setup. (I have tried this over a few times, with the red+blue button combo, doing the USB stick again. serial logs show this all seems to happen as expected). But after reaching the red light after install, on next boot we're just stuck back in the boot-loop
It looks like I can log into the serial console with root and fire off 1-2 commands before a forced reboot. Some ‘ps’ and docker logs later, I didn’t uncover anything obvious. Is there a command I can run to pause whatever watchdog or supervisor is calling everything off? Will be easier to poke around and find what’s failing
located the error, even if not very informative:
23-11-26 23:21:16 WARNING (MainThread) [supervisor.homeassistant.core] Watchdog found Home Assistant failed, restarting...
23-11-26 23:21:16 ERROR (MainThread) [supervisor.homeassistant.core] Watchdog restart of Home Assistant failed!
23-11-26 23:21:16 INFO (SyncWorker_0) [supervisor.docker.manager] Starting homeassistant
23-11-26 23:21:16 INFO (MainThread) [supervisor.os.manager] Detect Home Assistant Operating System 11.1 / BootSlot A
23-11-26 23:21:17 WARNING (SyncWorker_0) [supervisor.store.data] Can't read /data/addons/core/zwave_js/config.yaml: does not match regular expression ^(config|ssl|addons|backup|share|media)(?::(rw|ro))?$ @ data['map'][0]. Got 'addon_config:rw'
23-11-26 23:21:17 WARNING (SyncWorker_0) [supervisor.store.data] Can't read /data/addons/core/configurator/config.yaml: does not match regular expression ^(config|ssl|addons|backup|share|media)(?::(rw|ro))?$ @ data['map'][0]. Got 'all_addon_configs:rw'
does not match regular expression ^(config|ssl|addons|backup|share|media)(?::(rw|ro))?$ @ data['map'][2]. Got 'homeassistant_config:rw'
23-11-26 23:21:17 WARNING (SyncWorker_0) [supervisor.store.data] Can't read /data/addons/core/samba/config.yaml: does not match regular expression ^(config|ssl|addons|backup|share|media)(?::(rw|ro))?$ @ data['map'][1]. Got 'all_addon_configs:rw'
not sure if all the config regex warnings are important, or whats causing things to crash. This is just straight from install procedure, have not even been able to complete the webpage setup. So the config stuff is just however it's installing itself
last lines before reboot:
23-11-26 23:32:59 INFO (MainThread) [supervisor.supervisor] Fetching AppArmor profile https://version.home-assistant.io/apparmor_stable.txt
23-11-26 23:32:59 INFO (MainThread) [supervisor.host.apparmor] Adding/updating AppArmor profile: hassio-supervisor
23-11-26 23:32:59 INFO (MainThread) [supervisor.supervisor] Update Supervisor to version 2023.11.3
23-11-26 Stopping Serial Getty on ttyAMA2...
Stopping Hostname Service...
Stopping User Login Management...
Stopping Load/Save Random Seed...
Did you choose the “home assistant for yellow” image in rpi imager or one of the other home assistant images? It’s below the others.
using rpi-imager I did HA for yellow (selected RPi4 in device selector, then "other specific-purpose OS" > Home assistants and home automation > Home Assistant > Home Assistant OS Installer for Yellow).
letting it reboot over-and-over for ~8h did not fix. At one of the reboots it stopped and started network time sync service about 20 times, then got stuck waiting for.. a few services:
[ OK ] Started Network Time Synchronization.
[ OK ] Stopped Network Time Synchronization.
Starting Network Time Synchronization...
[ OK ] Started Network Time Synchronization.
[ *** ] (1 of 3) A start job is running for…tc/fstab (5h 58min 47s / no limit)
After waking up and checking on this, I powered down, took it all apart, and removed and re-seated the pi cm4 and the nvme. No change after re-seating
if anyone else is curious, I did finally get the reboot-loop to stop. Absent any errors, I believed the board was overheating. I checked the thermal pads and the heatsink were in secure positive contact. Not enough. Then, active cooling across the fins with a fan was not enough. I slowly dripped the contents of an upside-down compressed air canister onto the heatsink. That did work, boot finally finished, and I got to see the setup screen finally
🤨
it could be completely coincidental timing that it finally works on the one attempt I shoot compressed air at it. but after going through almost 2 days of reboots, it finally works when I freeze it? hard to draw any other conclusion
That doesn't sound healty
@wind saffron no that does not sound good. Do you happen to have a 12v 2A power supply you could try instead?
that is a great question, and probably something I should have also tried! unfortunately, (or, fortunately for me) this was only a problem during the setup stage of first boot post-install. Once I cooled it enough to get past that stage, it has not needed any active cooling since. So I can't (thankfully?) reproduce the issue anymore to try this.
Just to be sure - does it reboot (not just restart) without issues?
I can unplug and unpower it, and power it back up without issue right now, yes. With the case on
That's good. But still strange
I don't think so. The people in #cameras-archived might be more familiar with Scrypted's hardware requirements
2023-11-29 16:02:35.293 ERROR (MainThread) [homeassistant.setup] Unable to set up dependencies of esphom
e. Setup failed for dependencies: bluetooth
getting this over and over on a fresh install
In what context? What are you doing?
trying to run home assistant on a home assistant yellow lol
it's a fresh install and home assistant isn't coming up
looking through serial and that's the only error i was able to find
in ha core logs
Are you restoring a backup?
And it has been running on the same hardware before?
no, this is a brand new yellow
How did you install HA?
ha yellow installer with USB
trying installation process again watching serial console this time
Are you using a NVMe?
Has it been used before with HA?
no, it's also brand new
[ 31.759153] haos-flash[247]: Installing Home Assitant OS 11.1 to nvme0n1.
seems to be going fine
Do you have a CM4 with eMMC storage?
no, lite
Okay. So no storage where an old version could be.
oh well nvm
now i'm at the "preparing home assistant screen" when connecting over :8123
maybe i was just impatient or something on the last install lol
Wouldn't be the first time some strange hiccup happened 
it is what it is
oh also for the record for anybody like me who does a few searches on discord and forums before doing anything:
SK Hynix Platinum P41 does work with Home Assistant Yellow
Thanks for sharing that info!
so is continued heartbeat the expected behavior for the yellow LED if it's not configured to off?
When HAOS is running, the yellow LED should do the heartbeat pattern, as long as it is not turned off.
makes sense
||(and if you are not abusing it for something else
)||
Can I use any version of the Reapberry PI version 4 on a yellow? I'm considering getting myself a yellow and would be a lot cheaper to use the pi I have already
No. It needs to be a Raspberry Pi Compute Module 4
A ususal RPi 4 won't work with the Yellow
Thank you
I'm having trouble with my new Yellow that arrived today. I got it working at first, but then it seemed to stop connecting to my network so I decided to factory reset (since I hadn't gotten too far anyway). Now I'm at the stage where I have a solid yellow and solid red light. I understand this indicates a network issue, but the device is still showing up in my router, just with no activity. Also, I have the CM4 lite with an SSD installed in the NVMe slot. Any ideas?
Ok, I guess 3rd time's the charm... I don't understand what I did, but I finally made it back to the setup screen!
Hi,
i needed to revert to Homeassistant OS 10.5, since OS 11.x is not detecting my USB2Serial (Prolific) Controllers on my Yellow. What is the best way (what informations) to raise an bug in github?
Don't want to jump back and forth between the OS Versions a lot 😅
The requested information is stated in the issue form
https://github.com/home-assistant/operating-system/issues/new/choose
If anything more is needed, they will ask for it
Maybe more details about the USB device
folks, the random-reboot guy is back. I have tried installing two addons (one, the traditional SSH addon, the other is the esphome addon). Both failed, so now I'm troubleshooting that. Now, I can trigger the immediate shutdown when I just pull a docker image. The system just shuts down instead of finishing the command (also ha su repair makes it just shut off immediately). I have switched from powering with PoE to using a 12v 5A psu, so there should be plenty of overhead and capacity.
Use `ha` to access the Home Assistant CLI.
# docker pull ghcr.io/hassio-addons/ssh/aarch64:16.0.1
16.0.1: Pulling from hassio-addons/ssh/aarch64
579b34f0a95b: Already exists
27a08995f1f8: Extracting 524.3kB/11.05MB
e870c2d39f10: Download complete
00400efbd5cb: Download complete
bb322439d86b: Download complete
ebb9dca6e7d9: Downloading 12.43MB/91.87MB
76ddbf20a8c0: Download complete
87ff96e5be9e: Download complete
unexpected EOF
# [ OK ] Removed slice Slice /system/modprobe.
[ OK ] Stopped target Bluetooth Support.
[ OK ] Stopped target Multi-User System.
You already have tried to re-install the OS, right?
yes
If you have already checked, if the CM4 is fully seated in its sockets, you might want to contact the support at yellow@home-assistant.io , explain the situation and that they can contact you here. It might be a hardware problem.
Yeah, I did disassemble and re-seat the nvme and CM4 module, got a very authoritative click from both sides. I sent a support email during the boot-up issues, haven't heard back. Pointed to the channel here as well
Well, you heard back from the team. But it looked like you have been able to solve your issue.
@wind saffron is it getting hot again when pulling the images?
not noticeably, no. I did try the compressed-air thing again, without the same success. I'm not as convinced it was temp/overheating? I did get the images to pull successfully, if I timed the command about 2 minutes after boot login prompt (from serial). If I waited longer, system shut off. If I did it too early, system shut off
It seemed to coincide with a lull in other things using up CPU. Right at boot, a lot of things are starting up, etc. If I did it after all that quieted down (looking at top) I could clear the image pulling.
Whether that meant too much CPU activity = hot = same conclusion as before, possibly.
I went with a 12v 5a barrel jack power supply to make sure it wasn't a brown-out type of thing when the cpu got busy
I haven't seen many Sabrent Rockets to know if they might be an issue. That's what you're using, correct?
that is the nvme, yes
I would assume that's not directly related to increased CPU but @shut cedar might have better suggestions about what to check for
And then there's the Lite itself. This was new out of the box?
yes
Uh that is interesting. Is journalctl logginig any reason for the shutdown? 🤔
Installed 11.2 but my Yellow never came back up 😦
Console stuck like this
homeassistant login:
homeassistant login:
homeassistant login: root
Last login: Mon Dec 4 21:02:13 on ttyAMA2
Welcome to Home Assistant OS.
Use `ha` to access the Home Assistant CLI.
# ha
What does ha os info say?
Just got my POE Yellow and going through the install.
At the Pi Imager step it says
if you are using a Raspberry Pi Compute Module 4 (CM4) that already has software on it, this procedure does not work. Ignore this and the following steps.
Is there an easy way to know if my CM4 has software on it?
Have you been using the CM4 before or have you gotten it second hand?
No, just bought it in conjunction with the box/board
I assume it's been in the original Raspberry Pi packaging? In that case no software is on it
it just hangs there, only command I can get to return is top
Um, it just came in a little cardboard box, but I'm assuming so.
I'd try a reboot
[ OK ] Started Bluetooth service.
[ OK ] Reached target Bluetooth Support.
[ OK ] Finished HassOS Configuration Manager.
Starting Network Manager Script Dispatcher Service...
[ OK ] Started Network Manager Script Dispatcher Service.
[FAILED] Failed to start containerd container runtime.
See 'systemctl status containerd.service' for details.
[ OK ] Reached target Multi-User System.
Welcome to Home Assistant
homeassistant login:
this is where it gets to on a fresh boot
Login and try ha super repair
hm, just hangs there
Just to confirm, the yellow does not support wifi by default right? It's got built in Zigbee/Thread, but that's it?
If the CM4 has no built-in WiFi/BT, that's correct. If it has, it is supported, but not recommended (and not covered by the regulatory approval).
👍 My eventual solution in the new house will use a POE switch, but currently I'm using a spare POE injecter I've got, which means multiple ethernet cables, so I was wondering if I could do the rest of setup over wifi after the initial install, but it sounds like no.
I've found a few errors so far. Trying to export the journal logs from each boot so it's easier to sift through than USB serial. A common occurrence seems to be a golang-runtime panic coming out of some crypto modules when doing TLS. When the process quits, supervisor seems to bring everything else down with it. Then I just got this spit out:
@wind saffron I converted your message into a file since it's above 15 lines :+1:
although it doesn't seem to have crashed/restarted with that paging error.. it just spat that out over serial and continued on ok. The golang crypto errors definitely preceded the system reboot though
You can still use a 12V/2A power supply with an according barrel jack
Hm, that sounds like some type of read/disk communication issue 😢 What NVMe are you using?
@shut cedar a Sabrent rocket. Haven't seen many others who have tried using these
I'm looking for a USB stick cell modem 4G LTE that will allow me to send SMS from the Yellow that works in US AT&T
There are also integrations for Huawei LTE, TP-Link LTE or Netgear LTE. You would have to find some compatible hardware for one of those.
I've looked at that info and the Huawei devices are unavailable on Amazon that's why I asked for something known to work not on that list . How likely is something in the Gammu database likely to work? Additionally I don't have a good understanding of all different technologies so looking for something know to work with HA and on a Yellow.
I posted on #hardware-archived for available recommendations
Thanks
That Gammu list has some really old devices
@vestal token I have found something, what is widely available. Not nice looking, but should work for SMS and is quite affordable.
USB sticks with SIM800C chip.
https://www.amazon.com/s?k=SIM800C+usb
I'm aware of this device but I think the SIM800C its actually only 2G the wording in the description is misleading . later in the details they reference compatibility with 2G/3G/4G Sims, but 2G network. I know some people are using it. But as 2G is no longer officially supported it may not work for a carrier or in a specific area. I might try it because it's cheap and I can send it back if it doesn't work. Even if it does work, it could quit being supported at any time.
https://www.stlfinder.com/model/sim800c-case-32La84DY/4922132/ case for 3d print
3G is depreciated in favor for 4G in many areas. But afaik 2G stays to keep SMS and so on working.
I've heard that but it would be nice to see something official about support and EOL ? I've been hoping SIMCom would come out with a newer fully compatible similar device
Trying a newer Huawei LTE stick would be a gamble too. Its firmware might not be usable.
There are SIMCom 4G USB sticks. But much more expensive and no idea, if they can be used in HA.
By chance did you fix this by taking out the SSD? I am having the same issue where the Home Assistant Yellow will show normal operation LED light but not be able to connect to while an SSD was in the M.2 slot. For context I tried using a Microsoft Surface 8 256 gb SSD. The SSD controller is Marvell® 88SS1322 if there is any know issues for this as I have no clue.
No, maybe I just wasn’t patient enough or maybe my network cable wasn’t super stable (I’ve since had occasional issues where the device shows offline in my router).
Can anyone help me with this?
https://www.reddit.com/r/homeassistant/s/aR19Ir1CUa
I found some earlier comments mentioning the ssd. Looks like I bought a non nvme SSD... Just ordered a new one so let's hope that fixes this
How hard is it to get ZWave to work with the HA Yellow? Would I just buy a USB dongle? Is there a better way?
You can use a USB dongle or a Pi Z-Wave HAT like the Aeotec Z-Pi 7. The HAT would be attached to the GPIO pins within the casing. Both things aren't too hard
If I understand correctly I wouldn't be able to use bluetooth if I go the HAT approach, due to limitations with the Pi... so maybe a dongle would be better
That doesn't apply to the Yellow. You can still use both without altering any OS configuration
Oh, interesting
Wow, those HATs are pricy
Looks like if I went with the Z-Wave.Me RaZberry 7 Pro, I'd have to drill a hole in the case for the antenna... there's also the non-pro version, which I assume would just have a slightly lower range.
Z-Wave is a mesh network, so range shouldn't be a big issue, if routing devices are nearby
Ah, that makes sense
There's so much to learn, I didn't mean to go down a rabbit hole today >_<
My HA Yellow (using a nvme-based boot/install) routinely boots to the wrong boot image (so, for example, it's currently booting into HAOS 10.5 despite 11.1 being installed in the other slot). When this occurs, I can sometimes get it to flip back to the other slot, but it takes a good number of reboots and randomness. Is there a way for me to froce it to attempt to boot to a specific boot image on next boot? There's no obviosu errors in the boot look -- so I'm not sure what's triggering it to fail over to teh alternate slot. But once it fails over, getting it to retry the main slot is difficult. Any ideas?
@languid pulsar have you tried updating (while in 10.5) to 11.1 again?
I would expect it to re-write 11.1 fresh to the other slot, but not sure
I can do that, but it just keeps installing the latest version to the other slot. And the alternative slot stays stuck on 10.5. which works to some degree, but I'd like to get the alternate slot updated as well.
Is there a way to force an update on a given slot?
It should stay stuck in order to provide the fallback
I don't know of a way to force it
I want the fall back to be a newer version though.
so while in 11.1, updating to 11.2 should replace 10.5 on the other slot
All these versions work. I'm not actually sure why it fails over, but it seems random, not systemic so I'd like both slots to say least be within a version or two of each other.
It fails to boot 3 times, or is interrupted purposefully 3 times, it falls back
And I guess I should say, it falls back to the other slot. Not falls back to a lesser version. It would depend what is on each slot.
This is what I want, but because it's stuck booting to 10.5, 11.2 gets installed over 11.1 instead of over 10.5.
You can specify the version in the ssh terminal
What does that look like?
I'm now on 11.2 in one slot and 10.5 still in the other. Is there a way for me to update the 10.5 slot when booted into the 11.2 slot?
If you are actively in 11.2 when you update, that is how it should go
waiting for my reboot to complete where I am testing this
@languid pulsar do you have access to the ssh addon terminal?
ha os info there will show you the version of the os and the slot
I was on 11.2 in slot A
Yeah, I have SSH access.
I updated to 11.1 and it has rebooted in slot B
K, I'll give that a go.
Once I do that to update slot B to 11.1, do rerun the 11.2 update from slot B to get back to the 11.2 install in Slot A?
It would be nice if we could just set the next boot slot directly.
updates will always take up the other slot
and that includes "updating" to a lower version
But if you are having an issue where it randomly changes and you are not running the version you expect, I suspect some kind of corruption?
Any around that could help me out on a stuck install? I just got a new yellow and added a cm4 lite and a nVME and cannot get anything to install. I have check the cm4 and the nVME and they are both fine.
When powering up the red and green lights are on. After a few seconds the yellow goes to a heartbeat blink with the green is randomly flashing. Then after about 5-10 sec the yellow goes solid and a very faint green light on heartbeat. Not a network issues from what I can tell as I have a device showing up on the network called yellow-installer.lan. Any help would be great as I have been at this for 2 days.
@fluid hound You might try directly flashing the nVME SSD with the Yellow image of HAOS if the USB installer method isn't working
Ok, I will let the client know it will be a few more days. I will need to order an external kit to do that.
And what NVMe model are you using?
A Samsung 970 EVO Plus. I know they work as I have used these same drives in 4 other yellow setups. This is the only one yellow that will not work right.
If you're doing this on a regular basis you probably want an adapter anyway. Tools for the job and all that. Are you using Lites in the others as well?
Yes I am. All identical setups, just this one refuses to install.
Oh yeah on the enclosure. I am dumb for not having one in the first place.
If flashing the HAOS image for Yellow doesn't help, the output of the USB-C serial console might give some hints
I haven't been able to track down the underlying bug. What logs would show teh error leading to it to fail over to the other boot slot? The kernel boot log itself never seems to show anything, and I don't see any errors in the HaOS logs.
Does anyone know when to expect ameriDroid or Crowd supply stock for the US? Specifically for the poe model.
Hey, is pasting images in this channel disabled or do I just lack permissions as a new user?
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.
my nvme drive seems to have just completely died. are there any recommended debugging steps to see if I can get it to connect again? lsblk doesn’t show the device at all
Has it been running before with the Yellow?
yeah, has been running for 6 months. just died overnight
i was able to restore from a backup from 4am today. but can’t get that nvme drive recognized
luckily my backups go to a nas share
And you are using lsblk on the serial console?
Have you removed everything from the NVMe? Easiest way would be a m.2 to USB adapter and removing all existing partitions. Afaik HA doesn't show NVMes with a preexisting HA installation on it for data migration.
i tried an m2 to usb and my macbook won’t see it in disk utility either. i think it’s just dead
yeah, serial console
If another device doesn't see it too, it sounds to be toast, yes
ok. well things seem to be running again. will order a new one
Is there a known issue with the yellow shutting down when installing stuff from hacs?
It has happened twice today, and now i cant get it online
No. Working fine here.
ok thanks
Not sure what to do actually.
It shut down during installation of a hacs integration, now i cant get it online again.
Might it just take a while?
You can use the USB-C serial console to look for error messages.
Or look at the observer, if it is showing something
http://homeassistant.local:4357
its back on, just took a while.. is there a way to figure out why it shut down?
wierd.. it failed again.. but states are "heatly" on the observer
i got to the login page, and then it disconnected again.
You can look at the log files or use ha core logs and/or ha supervisor logs.
There is also the option to observe the boot process via the serial console
ok thank you
I got it back online.. but it looks like it is struggling every time i go into hacs menu. i might need to downgrade
Hi Everyone, planning to buy a yellow for use as a thread border router? how good is it as a thread border router
I haven't got any Thread devices yet. #thread-archived might be able to tell you, how good the HA Thread integration is meanwhile.
Can i power the yellow via the USB-C port ?
No
ok 😦
Hi 🙂
I have some issues with constant crashes, what log will show me the reason for the reboot?
I feel like most logs are cleared after a crash
You can look at the logfile in /config or use journalctl in the serial console to get the OS logs
Ty
What is the most reliable NVMe to use with the yellow? I have a PoE yellow with a sabrent rocket brand nvme now, but things are flaky. I want to test with a known-good hardware configuration to rule out the drive
I use Samsung and Western Digital for all my drives
I think I only have Samsung 980 pros on hand, which I think is the one not-working product. I’ll try a WD SN770
WD Black SN770 is reported as not compatible @wind saffron
Samsung 970 Evo Plus is a safe call, for example.
D’oh! I should have checked the list again. Thank you for the update
Hello everyone,
I am new to HA and hoping to buy HA Yellow.
I checked all the recommended sellers on the official website and came across this one - https://www.seeedstudio.com/Home-Assistant-Yellow-Standard-Version-with-CM4-p-5809.html
However it says - The version of CM4(CM4002016) on the board is not equipped with wireless ability.
Which model is recommended to buy? I am certain I want to buy HA Yellow Standard however dont want to miss out on the wireless ability.
Any recommendations?
You don't need the wireless option CM4. All wireless network devices can be accessed via the wired network connection of the Yellow. And for Bluetooth: you could still use a USB dongle, but an ESPHome Bluetooth proxy would be even better.
I'd go with a CM4 with more RAM and storage.
The Yellow Standard has been upgraded regarding RAM and storage for a while, but I don't know if that's still the case at Seeed Studio. I would suggest to ask their customer support before ordering.
I got it from Seeed (great company in my limited experience) with the 8GB CM4 and I can at least confirm that the Bluetooth works fine. I haven't tested the WiFi, because as Jorg says you'll ideally be going through your router/AP for any WiFi devices
I'm new here so I can't comment on the details because I'm still setting it up, but for the couple bucks extra Seeed wants, if you're worried about it, maybe bundle the external antenna in in case you want it later.
I got it because I do use my Yellow for ingesting data from my old smart scale over Bluetooth, and I wanted to make sure it had a bit of extra range
I haven't experienced much differences with a external PCB antenna (not the big one).
Yeah, I forgot to enable it, so I can't say yet, but I was able to drill through the case to mount it as some people suggested. It was my first time drilling acrylic, and I don't have a great drill, but I put electrical tape on both sides and used increasingly large drill bits. I guess there's no harm in it
I didn't want to drill, so I used a PCB antenna like this. Wasn't too impressed with the results.
ah. Oh my bad, I think it's polycarbonate. Either way 🤷♀️
My searching in here and elsewhere has returned no results, so here is my question/issue:
I bought a CM4 off of ebay during the pandemic for an eventual HA Yellow build, and I have finally purchased everything I need to do it. I am now running into the issue of having no way (to my knowledge) to overwrite the "Recalbox" installation that is already on the CM4 flash memory. I am using an NVME drive, I do not have a CM4 breakout board, and I am not sure what to do from here. Researching everything from "wipe/factory reset a CM4" to "HA yellow force boot from USB" and everything in between, to no avail. Not experienced with any of this technically, and looking for some help. Thanks!
Cross posted from installation
Tried this?
@quiet harbor yes I did, though I have a Mac and couldn’t complete the second option.
Make sure it's properly inserted in the socket
Many issues with wiping etc are caused by the cm4 not being properly inserted
https://www.jeffgeerling.com/blog/2020/flashing-raspberry-pi-compute-module-on-macos-usbboot here's a way to get rpiboot working on a mac for option 2. But perhaps @wary pivot has other advice
Yeah. Jeff is basically doing the Mac stuff from the usbboot docs
https://github.com/raspberrypi/usbboot/blob/master/Readme.md
if I have a CM4 lite in my yellow, and I am swapping nvme's, can I just clone from old to new (same size drives), put the new one in the yellow, and it'll work? Or is there a procedure or incantation to follow first?
I physically reinstalled the cm4. In the meantime, I got rpiboot on my wife’s work pc, and have still failed to successfully run it on the cm4/yellow. I keep getting a series of error messages.
Perfect! Thank you @wary pivot and @modern basin! Really appreciate your feedback.
I haven't tried to clone a drive yet, so not sure if it works. But chances are good. I would try that, as that might save a lot of time. But make a full backup and download it first, just in case. If it doesn't work, you can still reinstall HAOS and restore the backup.
Did you do this?
Advanced: Forcing installation onto the NVMe SSD when using CM4 with eMMC:
Press the blue button while the yellow LED is on constantly (during the 5s window, see next step).
hey BEEF did you ever get this sorted out? I been running thread on a HA yellow for almost a month now and it seems like every day i need to restart multiprotocol add on because thread network has shit the bed
Here we go
https://pasteboard.co/i97mGFMc9tFh.jpg
Cheers, I hope it goes well.
random-reboot on install guy is back. I switched out my nvme from a sabrent rocket to a samsung 970 evo plus. My Yellow continues to just.. reboot itself over and over during post-install. I've eliminated most hardware issues. I tried a 12v barrel plug instead of PoE power, I have now tried a known-good nvme. I re-seated the CM4 yet again (paying VERY close attention to a good click from each side). Something here is a dud and I can't figure it out
the upside-down compressed air-as-coolant trick is not working this time
Dec 16 19:26:49 homeassistant hassio_supervisor[537]: 23-12-16 14:26:49 INFO (MainThread) [supervisor.resolution.fixup] Starting system autofix at state running
Dec 16 19:26:49 homeassistant hassio_supervisor[537]: 23-12-16 14:26:49 INFO (MainThread) [supervisor.resolution.fixup] System autofix complete
Dec 16 19:26:50 homeassistant hassio_observer[537]: 2023/12/16 19:26:50 Access to logs from 172.30.32.1:51524
Dec 16 19:26:52 homeassistant dockerd[537]: unexpected fault address 0x8ce6328b98399a6c
Dec 16 19:26:52 homeassistant dockerd[537]: fatal error: fault
Dec 16 19:26:52 homeassistant dockerd[537]: [signal SIGSEGV: segmentation violation code=0x1 addr=0x8ce6328b98399a6c pc=0x47128c]
Dec 16 19:26:52 homeassistant dockerd[537]: goroutine 1140 [running]:
to the best of my golang-callstack interpretation, it looks like docker is having a segfault while doing a docker pull.
why it's crashing, I can't figure out
also why I'm the only person who seems to be stuck here I can't figure out
this could also explain why I couldn't install any addons when I did finally get things to work: it was probably pulling docker images, which seems to trigger the crash
more logs: https://dpaste.org/mdnD9
its also very difficult to get supervisor logs: https://dpaste.org/QG3kd
How much RAM does your CM4 have?
As often as you have installed HAOS (which is running fine here), I would say a software issue is not too likely. My guess would be a hardware defect. If I had to place a bet, I would say something is wrong with the RAM. What do you think, @compact quartz and @shut cedar ?
it's a CM4 8GB lite
from the last time I went through this, If i figure out what image it is trying to pull, and run docker pull **** from the serial terminal, I can make some 30% or so progress before reboot, and after a few reboots, if I get it to finish.. install continues to the next stage and next container image. Same thing happened when I tried to add the SSH addon, immediate crash when it tried to pull the docker image, but if i serial console in and pull it ahead of time, I can get things to continue
once I manually pull the images to get it to boot, I will try downloading and running memtester to check the Pi's ram
so I ran a cpu bench, to test any problems with power draw or overheating -- 20 minutes of redlined CPU didn't trigger any reboots. Then I ran memtester (granted I had to also do this from within a container, docker exec -it homeassistant /bin/bash then apk add memtester and run it). I have not produced any RAM errors from memtester as of yet. But I will run a large loop for the next few hours
2 hours straight of memtester, no issues so far
But if the OS tries to docker pull anything, I get an immediate golang crash, stack trace all about the network operations, etc. weird! I cannot get a golang panic when issuing pulls from the serial console
@wind saffron Tried a different SSD?
yes, I've got a samsung 970 evo plus in there now
ok, its absolutely a docker and/or network thing now. I ran cpu bench for half an hour, all fine. Ran memtester for 4 hours continuously. I finally just ctrl-c that, and ran a single docker pull python-alpine (as a random test image) and the pi immediately shut down
not even sure how to investigate this
at least yours boots.. after install mine comes up to nothing at all. it gets an DHCP IP, but has no serial output and shows no network activity (no ping, no open ports)
anyone know if this is common? I can try to reseat everything. it's a SK Hynix SSD out of an old Dell laptop. I have a Samsung 960 EVO I can try.
At least the SK Hynix BC711 is reported as incompatible. Reseating the CM4 is a good idea, if changing the NVMe doesn't work.
ok brb
it does blow my mind how a standards based device can be incompatible though
what's the hang up there
Not everyone follows the standards 100%. And sometimes it is a hardware bug, like the incompatibility with the m.2 Google Coral - which is caused by a CM4 hardware bug.
yeah I've read through that thread. that was pretty unfortunate.
do I need to delete all the partitions on the SSD first? or will the installer do it?
If those are HA partitions, you need to remove them. Otherwise it shouldn't matter
(But removing them wouldn't hurt)
it worked! used a WD Black I had.
the SK disk is a PC300 model, but that could just be an OEM version of the BC711, who knows
You are lucky. Some WD Blacks are repored as incompatible too 😆
SN750 SE, SN770 and SN850X
yeah I thought I remembered it saying that but I didn't want to waste the 500 GB Samsung so I gave it a shot. it's an older one, SN700
250 GB
Anyone else here using matter and thread on the yellow with multiprotocol radio? My thread network fails almost everyday requiring a restart of the silicon labs addon
Also watchdog is on so it's not crashing hard... Just getting into a weird state
i won't have any thread devices until next week
How much power is the yellow PoE using when running the HA?
I have a solid red light and a heartbeat yellow light happening. I had HA running for a few weeks and it was great, but I had a power outage and I cannot access my HA with either the app or the browser.
The indicator lights make it seem like everything is working fine so I am very confused. Any help is welcome
So I just searched to add a new server and it found my HA and all the devices on it... Why did that change ip?
If your IP hasn't been set as static in HA or hasn't been reserved in your router, the DHCP lease might have been expired.
Slightly worse would be a data corruption due to the ungraceful poweroff
Ooh thanks for the heads up. I'll have to look into making is a static ip
I'm looking to switch my yellow to use postgres as a DB instead of sqlite3 (recorder/history plugins have been failing to start up because of the amount of history & data, I have energy sensors that send a lot, and frequently). What's the recommended way to do that on Yellow specifically? IIRC the OS situation is somewhat immutable and services mostly managed through docker. I can't exactly apk add over the serial console
One caveat is that I'm still the guy who's Yellow crashes whenever I try to install an addon. So I do not have the SSH addon. I can only serial console over USB until I figure out the panics coming from docker pull
as in, how should I install and configure it as a dependent service in a Yellow environment. I can deal with the configs and DB admin stuff after that, just looking for advice for install and service config for this OS
HAOS is for big parts a read-only system. Installing software on it will most likely not work.
understood, so is there a path forward to using mysql/maria/postgres instead of sqlite on the yellow?
The crashes when using docker are going to have the only priority here I expect.
Since HAOS uses docker
How is updating HA going to work
if I docker pull manually over serial, I can get like.. 20% of progress done, then it reboots. If I spend an hour or so doing it over and over again, I can get it to finish. That's the only way I got install to complete
so if that's the way to get it going, I'll spend the hour and do it, but wanted to see if there was an alternative
All possible network parts that can be swapped have been swapped? Ethernet cable, ports on switch, or avoiding a switch in use?
I do plan on finding a long-term fix to why this doesn't work, whether it means proxying docker requests, or local-caching images I need to pull, hell even a PR to the docker codebase if it turns out to be an arm64 platform bug in the go stack, sure
but I'm already having history problems, and my graphs either stall or the history integration won't start up, so I'm fixing the DB thing first
Yeah I've tried the network hardware troubleshooting bit. Used a different cable, I took the PoE injector out of the loop so it was non-poe ethernet direct to the switch and used the barrel plug. Tried more different cables, shielded metal plugs, and non, different ports, etc
I'd suspect a hardware issue of some kind here. If you get random crashes all over the place, things are not good.
It could be memory, CPU (SoC) or maybe power (supply).
not so much random, as it is specifically a golang panic, at Sha256.Write() when I do docker pull
its a pretty consistent stack trace
I think the Raspberry Pi Power Supply Checker integration in Home Assistant works. Can you enable that and see if it reports "OK"? Then maybe pull another image, and see if it remains OK for the time.
Hm, but then maybe HAOS/that code i corrupted?
Did you try downgrade/upgrade HAOS (which will rewrite the squashfs rootfs)
I have not tried other HAOS versions. My install method was rpi-imager onto USB stick, I plug the stick in and it does its stuff, then next boot the yellow pulls its own docker images. I'm not sure how I would install other versions
this is the stack trace, if anyone was curious: https://pastebin.com/JPZwcN41
which, depending on the right version, is a panic at here: https://github.com/golang/go/blob/master/src/crypto/sha256/sha256.go#L198
What version do you have now?
Core 2023.11.3
Supervisor 2023.11.6
Operating System 11.1
Frontend 20231030.2
Ok, so at the console use ha os update --version=11.2
lmao, I can try that certainly. I don't have high hopes unless we think there is something in that release that fixes a bug related to this. I'm not sure we're looking at corrupt data/files here, but I am absolutely willing to try and rule it out
It will take me a few hours to get this to finish, I imagine, but I will do it and be back later
This will use the next boot slot, and then HAOS boots from that slot
Assuming it doesn't crash while downloading
given the restarts during pulls, surely something could be incorrectly pieced together through there, but I believe there are hash integrity checks of docker images
This isn't a docker pull but that might not matter
# ha os update --version=11.2
Error: 'OSManager.update' blocked from execution, supervisor needs to be updated first
ha su update first?
Just to be clear: As per analtyics, we have thousands of Yellow installations (https://analytics.home-assistant.io/). I myself run a Yellow in production and have a test unit here. This is not a general software bug. This is something specific to your environment/hardware.
Now as to what/why that happens in your case, I am not sure. Since it is quite reproducible, I'd say some corrupted sectors on the eMMC might be a theory.
Ugh, taht needs a docker pull first. Hold on, we might be able to work around that.
this is a CM4 lite, and this is reproducable across two different NVMes (right now, a samsung 970 evo plus)
Are you on the serial console?
yes, usb serial
Ok that rules out eMMC obviously 😅 This is curious. Then maybe a NVMe communication issue? Do you have the same error/stack trace on different NVMe's?
I think it is a software bug, but in the network stack. I wouldn't be surprised if, say, the first TCP response packet in my network has a super large latency, or maybe just headers but no body, so the golang lib expecting to unpack the tarball was opened, but no actual data yet, so there's a panic
something like that
Ok, let's do a manual RAUC update then:
cd /mnt/data
curl -O -L https://github.com/home-assistant/operating-system/releases/download/11.2/haos_yellow-11.2.raucb
rauc install haos_yellow-11.2.raucb
...
reboot
I think the peculiarity of my environment is probably in the network
But that leading a system reboot? 🤔
Ok, check sha256 then after curl 😅
$ sha256sum haos_yellow-11.2.raucb
4ec4f5cfe24bac9e33e454a5d38a7998672bb8edb6c8eafa1df9aa39250adff7 haos_yellow-11.2.raucb
during OS install that was easy to explain: a systemd-required process was exiting with a golang panic, so systemd thinks a service failed, dependencies not met, systemd restarts the OS. It is not a hard crash, systemd goes through it's whole stopping services gracefully
why I can make systemd stop the world from the serial terminal, I don't know
Ah ok, yeah that make sense. We consider docker service system critical.
surely, then from the terminal I can probably crash the docker daemon when I pull
which systemd notices and restarts
# curl -O -L https://github.com/home-assistant/operating-system/releases/downloa
d/11.2/haos_yellow-11.2.raucb
% Total % Received % Xferd Average Speed Time Time Time Current
Dload Upload Total Spent Left Speed
0 0 0 0 0 0 0 0 --:--:-- --:--:-- --:--:-- 0
100 118M 100 118M 0 0 22.3M 0 0:00:05 0:00:05 --:--:-- 23.8M
# sha256sum haos_yellow-11.2.raucb
4ec4f5cfe24bac9e33e454a5d38a7998672bb8edb6c8eafa1df9aa39250adff7 haos_yellow-11.2.raucb
#
looking good so far
Docker should not choke when some corrupted network packages come in to the point it crashes. That would be a Denial of service vulnerability... But "should not" doesn't mean it is impossible 😅
the positive thing is that if this somehow is in the golang stack on their side, my day job is a software eng writing go, so I can track it down
the rauc install succeeded
From my experience with HAOS, disk/file corruption happen quite often end in go traces. This is probably mainly because HAOS is mostly just go software (as in, Docker/containerd are large go binaries).
Cool, so after reboot you are on a new rootfs squashfs (you can check using rauc status before and after the reboot)
ok, so I booted into 11.2, and then crashed it by running ha os version 😅
@wind saffron I converted your message into a file since it's above 15 lines :+1:
How much RAM with this Lite?
Ugh 😢
8gb ram
Yeah I'd suspect the CM4 here somehow... Do you happen to have the IO board and or a second CM4?
I actually have to go, I am sorry.
unfortunately that I do not
no prob, thanks for your time 👋
FWIW, something post-update must be pulling images because now it is in a reboot loop
Dec 19 15:53:13 homeassistant hassio_supervisor[538]: 23-12-19 10:53:13 INFO (MainThread) [supervisor.supervisor] Update Supervisor to version 2023.12.0
Dec 19 15:53:13 homeassistant hassio_supervisor[538]: 23-12-19 10:53:13 INFO (MainThread) [supervisor.docker.interface] Downloading docker image ghcr.io/home-assistant/aarch64-hassio-supervisor with tag 2023.12.0.
there it is
it is perhaps not the network. when pulling the yellow-homeassistant image, I can pull a 300MB layer just fine every boot, but as soon as it moves to the "extracting" step, the crash happens. speaks more to disk than network given that. though I'm on a 970 evo now. I ordered a new CM4, we will see in a week or so whether that helps /shrug
Hi, bought Home Assistant Yellow about a month ago and I am trying to set up a Nexa plug(433MWz Tellstick device). Do I need any external hardware or does the Home Assistant Yellow have what i need to communicate via 433 MWz?
433/866 MHz are not included. You would need additional hardware.
And likely additional software/add-ons
thanks so much, I think I actually have the software/add-ons that i need
didnt realise i needed additional hardware
@wind saffron did you order a CM4 with emmc?
no. I can see how that could rule out the nvme issue if it behaves differently via emmc storage, but it's another lite
latest core update breaks my zigbee network on my yellow
im using zha and its definately the update as it works fine if I role it back
Everything seems to become unavailable for a few seconds at random, working fine 1 minute unavailable the next any ideas please?
No ZHA problems with 2023.12.3 here.
no offence but just because we have the same car doesnt mean mine cant be broken if yours isnt
its fine ill figure it out or ill not update for the moment its not a big deal but seems wierd only happened when i updated
Good luck 
You should update again and share the logs, but it sounds like it is not related to the update or the Yellow.
if I role back it works fine if I update it breaks how can that sound like its not related
Can you upload a ZHA debug log?
@vivid grove That's what I'm saying. Update, break it, share the logs
how do i log just the zha if i turn on debugging the log is huge and contains everything
That's fine. If you can, just ZIP the file up. You can share it either directly with a DM or upload it to a file sharing site (e.g. iCloud, Google Drive, etc.) if it's too big
not really comfortable with sharing the logs in full I have no idea what security risks that might pose
I work on both ZHA and the firmware for your Yellow's Zigbee radio, in addition to being an employee of Nabu Casa, but you can redact whatever info you feel necessary
but the issue there is I have no idea what to redact since I have no idea whats exposed but a quick look and I see Amz-Security-Token=IQoJb........... for example
not trying to be awkward here just trying to be careful
The issue is that everything on the zigbee network becomes unavailable all at the same time for at least 30 seconds
its pretty random from what i can see but still fairly regular and this definately was no an issue in the previous release
There are a few issues being tracked but without logs it's not possible to say why specifically it's happening for you (or even if your issue is one of the few currently identified)
You can wait until the Thursday patch release to see if it's fixed but seeing what the logs have would be the best way to know what specific issue you have
Yeah thats the plan
I can live without the update for the moment and if I cant limit the logs to just zha then ill have to do that
Oh. No, you can limit them. Just click the "Enable debug logging" button on the ZHA integration page, let it run for an hour so it crashes, then click the button again to disable logging and download a log. It'll include warnings and errors from other integrations but it won't include debug logs from them.
yeah i did that but thats where i see things im not sure I should be sharing like the amz security token
it doesnt take an hour for it to crash its pretty regular and often it may even be regular I just havent timed it
think i found how to limit the debug to just the zigbee/zha
ok so it happened 6 minutes ago, 5 minutes ago, 2 minutes ago and about 10 seconds ago, so yeah its pretty random but happening a lot
Do I need a lutron hub if I buy their lutron smart light switch? I have both zigbee (from the yellow) and z-wave (via a USB dongle) available to me.
@weary hill if it's a Z-Wave device you should be fine with your own Z-Wave controller
I don't think lutron has any Z-Wave switches?
@compact quartz well that's where I am confused. I thought by getting my own server (HA Yellow) that I wouldn't need to buy any hubs. But the documentation is showing the lutron devices use their own protocol? I guess I am confused when I HAVE to use a hub, and when I dont have to
@weary hill lutron is proprietary yeah
Zwave and Zigbee are protocols which aren't locked to any one hub. So you don't need Ikea's hub for IKEA things, smartthings hub for smartthings things, etc if youve got your own Zigbee and Z-Wave controllers
I looked into the FAQ, and I am trying to search in the channel, but I can't find if the root cause for some NVMEs to not work has been found? so I am sorry if I missed the answer somewhere, but I am now asking here 🙂
Best resource for information about this would be https://github.com/home-assistant/operating-system/discussions/2235 .
But for explaining the technical details (as far as they are discovered yet) someone more qualified would have to take over.
thanks!!
oh and I found this thread also: https://github.com/raspberrypi/linux/issues/5217
this has plenty of details, and the last one notes that some NVMe require a clock signal to function. ok, that's plenty of info for me 🙂
If the device has Z-Wave marking, it's been tested and certified compliant with the Z-Wave standard and will work with a certified Z-Wave controller. Zigbee marked devices may or may not be 100% compliant and they are not required to be tested and certified. I use both. Z-Wave devices tend to be more expensive but generally are more reliable and more consistent performers. I've never had and issue with a Z-Wave device and I have some that are 8 years old. I use the Zooz 800 Z-Wave board in my Yellow and it's been 100% reliable. I've never had a Z-Wave device go AWOL. Zigbee devices can go unavailable for no apparent reason. I have an automation that checks for those Zigbee devices that go AWOL. Zigbee shares RF space with 2.4 GHz Wi-fi which can result in interference. Z-Wave uses 908 Mhz in the US and 868 in EU so Wi-fi- does not interfere.
in case anyone is following my saga, or for posterity if anyone else comes across this and has the same problem? I have narrowed my situation to reproducable kernel panics. I got one or two when running memtester from the root OS, not inside a container. But not consistently, nor could I definitely find a physical address space that was faulty. But on the docker front, I was able to fetch an image using ctr image pull <X> to download, then export to an image.tar, and now docker load -i image.tar does the traditional golang panic every single time. Still don't know whether it's RAM or hard drive, or something else. But it's a lot of kernel panics manifesting as page faults. I have two CM4s on order: an identical lite, and one with emmc. Will see what happens next week or so
the cm4 lite arrived way earlier than expected (way to go newark?). loaded the docker image first time no issue. I installed an addon from the UI with no issue (could never do that before, the image pull would crash the world). Appears my issue is resolved: was a faulty CM4.
installed all updates, and a few addons (both of those were impossible tasks before). Definitely resolved. Thank you all for your help and patience
wow that did come fast. Glad to hear it works as expected with a different CM4. Strange.
Glad you were able to resolved it! Thanks for the update!
# ha host logs
[ 125.455782] nvme0n1: I/O Cmd(0x2) @ LBA 130318240, 168 blocks, I/O Error (sct 0x3 / sc 0x71)
[ 125.464668] I/O error, dev nvme0n1, sector 130318240 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 2
[ 125.474242] nvme0n1: I/O Cmd(0x2) @ LBA 26595848, 256 blocks, I/O Error (sct 0x3 / sc 0x71)
[ 125.482695] I/O error, dev nvme0n1, sector 26595848 op 0x0:(READ) flags 0x80700 phys_seg 32 prio class 2
[ 125.492304] nvme0n1: I/O Cmd(0x2) @ LBA 222613632, 8 blocks, I/O Error (sct 0x3 / sc 0x71)
[ 125.500668] I/O error, dev nvme0n1, sector 222613632 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 2
[ 125.510179] nvme0n1: I/O Cmd(0x2) @ LBA 40676440, 240 blocks, I/O Error (sct 0x3 / sc 0x71)
[ 125.518660] I/O error, dev nvme0n1, sector 40676440 op 0x0:(READ) flags 0x80700 phys_seg 30 prio class 2
[ 125.528793] nvme0n1: I/O Cmd(0x2) @ LBA 15554600, 40 blocks, I/O Error (sct 0x3 / sc 0x71)
[ 125.537157] I/O error, dev nvme0n1, sector 15554600 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 2
[ 125.546573] nvme0n1: I/O Cmd(0x2) @ LBA 109398056, 8 blocks, I/O Error (sct 0x3 / sc 0x71)
[ 125.554949] I/O error, dev nvme0n1, sector 109398056 op 0x0:(READ) flags 0x80700 phys_seg 1 prio class 2
guessing this means faulty nvme?
Did you have a power outage or cut the power without shutting HA down?
core & all add-ons stopped responding so I would have cut the power then
If you access the host OS via the USB-C serial console, you can run nvme smart-log /dev/nvme0 -H to see the SMART data of the NVMe (the health data) or nvme error-log /dev/nvme -e 5 to see the last 5 NVMe error messages.
Hm... can I check if the Yellow sitting in front of me on the desk has PoE?
The label on the bottom of the device or packaging are mentioning PoE if it is capable to use it.
It also has additional electronic components here
The white product label on the bottom shows a model number. The PoE model numbers end with -POE
Thank you :)
can we fix the double negative with this FAQ item?
All M.2 SATA SSDs are NOT supported. Besides the following NVMe SSDs are known to NOT work:
I'm unsure if the list is good or bad, 'besides' and 'not' do not help clarify it when used together.
All SATAs do not work. Additionally to that the later mentioned NVMes are known to not work. I see no double negative.
maybe add a comma after Besides
or just remove besides, it clarifies it self, it is superflous
Is it okay to ask for assistance on here? I have what seems to be a non responsive HA Yellow which I’ve tried to factory reset multiple times to no avail
Sure. What type of Yellow do you have, which CM4 is attached and do you use a NVMe (if yes, which one exactly)? A good first step would be to connect to the USB-C serial console to look for error messages.
https://yellow.home-assistant.io/guides/use-serial-console-windows/
https://yellow.home-assistant.io/guides/use-serial-console-linux-macos/
👋 ☕ Morning! Cross posting here as I'm not sure the best channel: #zigbee-archived message
TLDR: Got my Yellow yesterday, and trying to pair my first Zigbee device: Aqara Mini Switch. But I'm not seeing it show up when trying to add the Zigbee Device 😿 .
@solemn fractal Please do not cross post. Read the channel description, post it and wait for folks to respond. Crossposting wastes people's time as they're unaware of the help you're getting elsewhere.
If you don't get any responses after an hour or more, and your message is no longer on screen, it is fine to re-post or post a link to it.
🫡 🙇♂️
uart driver
I thought it was a deep commentary on the nature of bits and their transmission through a serial interface
If it was a deep thought, if might lead to 42 🤔
Thanks. Regular powered version and no nvme
does Home Assistant Yellow have a serial number anywhere? I'm documenting all of my electronics and would like to note down any unique identifier i can
already have the MAC address documented
It should be on the sticker at the bottom.
Maybe a stupid question but do I need the Skyconnect USB stick if I plan to buy a HA Yellow?
No. The Yellow has a Zigbee radio integrated (the same as in the SkyConnect)
The SkyConnect would be "needed" if you plan to use Zigbee and Thread with two seperate radios. In theory it would also work with one radio and the Multiprotocol add-on, but that is not fit for production use yet
thanks for the quick reply. I am experienced in working with RaspberryPis and, for example, flashing images, etc., but these smart home standards are completely new territory for me.
What would be needed to also support Matter ans Z-Wave?
For Matter nothing is needed. If you want to use Matter over Thread and you need a second radio (like the SkyConnect) or the Multiprotocol add-on (not recommended).
For Z-Wave a Z-Wave radio (compatible to your country) is needed. There are USB dongles and Pi-HATs like the Aeotec Z-Pi 7. The HAT can be placed on the GPIO pins inside the case.
thanks again, that has answered some questions. i'm planning to buy a Home assistant Yellow(So no Pihat is needed) could you perhaps point me into the right direction for an z-Wave Radio in(for) the German Region?
#zwave-archived can help with radio recommendations. I am using the HAT Aeotec Z-Pi 7 inside the case.
When you say inside the case you mean inside the HA Yellow Case?
Yes
oh ok, my mistake. since the term HAT was mentioned, I was thinking of an installation on an RPi with an HAT.
thanks a lot...and have a good start into 2024 👍
ah...ok. I think this will be it.
I just realized I messed up and got myself a Model B. Any way to make it work with yellow or do I need to go and buy only the compute module?
It needs to be a Compute Module 4
Before I mess up again. @wary pivot would a CM4108032 (CM4, WiFi, 8 GB ram, 32 GB eMMC) work or should I be looking into getting some other variation?
That's fine. WiFi isn't really needed, so a version without would be fine too.
Perfect, thank you
Hi. I`m struggeling with my HD yellow beeing slow.. so i am looking for a upgrade.
I se your comment @wary pivot about the wifi not beeing needed, so i was thinking of a CM4008000 – Raspberry Pi Compute Module 4 – 8GB RAM – 0GB Storage (Lite) ?
I am using a 200gb SSD on the yellow card, so i gues a lite version is fine?
Yes
If you want to use Bluetooth devices, you can still add one of the recommended BT sticks or preferebly an ESPHome BT Proxy
thanks, i will order one right away! hoping the installation of the new compute module is to dificult for a noobie 😛
If I change the computer module, but OS is installed on the SSD, do I need to re install the HA again ?
If nothing is installed on the eMMC, it should work without reinstalling
If the OS is on the eMMC and the data partition moved to the NVMe, a reinstall would be needed
Noted thanks
Is there a way to roll back HA to the previous version, or to a specific version like 2023.12? There is currently an integration that is not working for me (shelly) and I'd like to have it in a working state until it's fixed. There is a ticket working on the issue now, and rather than ask for an ETA, I'd rather try to be patient while maintaining a working system.
Yes. If you use a terminal, SSH or the serial console you can run ha core update --version 2023.12.4

is HA yellow consider a server? i'm new to HA
Everyhing providing software services to other devices is a server
Changing disk to SAMSUNG 970 EVO PLUS M.2 NVME SSD 500 GT MZV7S500BW really made the difference on RP4 lite install 😄
Hey guys. I'm having some issues getting my HA Yellow that I've just received in the mail to see USB devices. I've connected over serial and had a look at the output from lsusb and dmesg, but there's nothing there that suggests to me that the OS is even seeing the known-good USB devices that I'm plugging in. Does anyone have any debugging suggestions or is it likely that I have faulty hardware? The USB ports do provide power, just the OS doesn't see anything plugged in.
The OS sometimes needs a full reboot to recognize new USB devices
I've tried multiple reboots, same result.
You might want to check, if the CM4 is fully seated in its sockets.
That's a usual reason for malfunctions.
What are the usb devices? Zwave controller? Bluetooth?
Bluetooth USB, it's a sabrent one on the list of compatible bluetooth USBs. I also tried just inserting a USB memory stick to try to see if it was some quirk with that specific device, but it didn't see that either. Also tried inserting the USBs to test and they show up fine in lsusb on my arch installation.
I'll try that now, thanks.
Are you certain that it’s possible to reseat the CM4 if I bought the yellow with the computer module preassembled? It doesn’t want to separate from the main board after I remove the screws, and I don’t want to break anything.
Okay sorry I see it’s possible.
Let me try a bit harder.
If you got a preassebled one (and it's not one of the first batches produced), the CM4 should be fine
I did. I can’t attach photos now but I’ve removed the heat sink, just the CM4 doesn’t want to separate from the main board and I don’t want to force it.
It was bought about three weeks ago.
You don't need to remove the CM4 for now
Alright. I’ll put it back together. Any ideas for further troubleshooting?
Depending on how many devices are connected to the USB port (and how powerhungry they are) a powered USB hub might be needed.
But if you plugged the BT dongle directly into the USB port (without a further passive USB hub), that shouldn't be the case. 🤔
It’s just a single USB device plugged into the port in each case. No hub or anything like that. I don’t think power requirements are the issue.
If there were power issues I’d also expect it to show up in the kernel buffer.
anyone getting a SiLabs Multiprocol add-on starting error on 11.3? i just upgrading OS and HA and getting AttributeError: module 'gpiod.chip' has no attribute 'OPEN_BY_NUMBER' trying to start the multiprotocol add-on but not sure if HA or OS is culprit. any particular order i should try reverting to find out which one?
You can revert to the previous version or wait for the new one to be released in a bit
ok. does a bit mean sometime today? and by that it sounds like this is a known issue?
Probably yes and yes: https://github.com/home-assistant/addons/issues/3391. I'm just waiting on a build to finish to test the fix.
ah there it is. i was searching around and couldn't find any. thanks for info i'll wait and look out for it
Update: tried it on another home assistant yellow (a friend ordered one two weeks after me) and it worked fine, so I'm thinking this is a hardware issue. Will do a factory reset just in case, but otherwise 99% sure it's hardware.
If you want to try disassembling the Yellow to try and re-seat the CM4, give it a shot. The USB controller pins are right in this corner: https://i.imgur.com/dL5CCzk.png
Just did this, no effect sadly. Thanks for the diagram though, appreciate the effort.
In the Yellow's Hardware page, click on "All Hardware". What shows up when you search for 1D6B? What about 1a40?
@bronze holly I converted your message into a file since it's above 15 lines :+1:
I haven't had issues detecting the USB hubs themselves, just getting them to detect anything inserted into them.
@bronze holly I converted your message into a file since it's above 15 lines :+1:
The Yellow has two USB hubs: the root hub, in the pi, and a secondary one for the two USB ports in the back. If the Yellow's CM4 is unseated, the two ports on the back basically are unplugged. The log sort of suggests the second hub is functional. When you re-seated the CM4, how did you do so?
Followed these instructions: https://yellow.home-assistant.io/guides/remove-cm4/
Then put it back in a similar way. Did my best to make sure it all made contact.
Behaviour before and after reseating seems the same. The home assistant/alpine installation detected the USB hubs just fine both before and after, just had never detected any device plugged into them.
I have an HA yellow that I would like to run WiFi only. I know you're not supposed to do this, but the location is in a home in a rural area where the only internet is a cellular hotspot. I am trying to install the yellow over WiFi (where I have GREAT WiFi) and will relocate it later to the rural location. I would prefer to never hook an ethernet cable to the box (challenge accepted). I tried creating the /config/network/ directory on the install USB with a WiFi config file for my netnet work in it...but it is not working. It has to be possible, right? I should mention that I have a PI CM with WiFi.
This might help you. You can access the CLI via the USB-C serial console.
https://community.home-assistant.io/t/guide-connecting-pi-with-home-assistant-os-to-wifi-or-other-networking-changes/98768
It might also work with the ha network commands using the CLI
Hi, quick question. I got a blue led blinking on my Home Assistant Yellow on the left side of it, is that the zigbee module?
Yes
Hi, a question regarding replacement of CM4. I ordered a more powerful CM4. Bascially i have understand the Docu how to install a new CM4. I replace the module and install the OS from a USB drive attached to the yellow to flash the eMMC. Once HA starts i can restore my backup during the on-boarding. My data is stored on SSD. Am i right that my data partition is immediately detected once i finished the restore?
Did you use a CM4 Lite before? In that case it would work without reinstallation (and not using the new CM4's eMMC). If the old one had eMMC (and if it was in use - not a NVMe-only install), you need to wipe the NVMe and move the data partition after install again.
The restore process should be improved, but I'd recommend to create an uncompressed backup using a service call in the developer tools. Those are bigger but seem to restore quicker.
I am using currently a CM4 with 2GB memory without wlan and bluetooth and attached internally a SSD an moved the data to the SSD.
The best way would be an uncompressed backup, downloading it and wiping all existing partitions on the NVMe
Wiping all patitions on the NVMe means the flash will erased? And does the uncompressed backup contains all data including the data stored on the NVMe?
All data on the NVMe will be kind of erased, when deleting the partitions. The uncompressed backup contains the same data as the backup via the UI. It should also contain all data in /config, /share, /media and so on.
If you worry about something special, you can open the backups (without password) with 7Zip for example and inspect it.
If it is with password it would need some decryption before you could open it.
Sorry but it is still not clear for me. If i restore a HA Backup it includes also the data stored on the NVMe? I am wondering why i must wipe the NVMe? I thought to store the data on a separated storage splitted from HA itself is one of the big advantages besides the performance. Therefore i do not understand why i should wipe the NVMe and i am fearing to lose data!
The backup includes it. The new install doesn't know it should mount the existing data partition on the NVMe. And with an existing old data partition, the NVMe might not be shown for data migration (which would delete the data anyway).
Hmm ok, i thought when i restore the backup HA knows or gets the information to use the NVMe and just continues uisng the data on the drive
Is it possible to create an overview which data is stored on the NVMe?
It is everything, except the OS itself
All configuration, custom integrations and other persistant directories are on the NVMe and backed up
Ok, understood. Can you give me a tip please how i can wipe the NVMe? Its a Samsung 970 EVO
I used a NVMe to USB adapter and the Windows drive management.
But it should also be possible to use the USB-C and usbboot to mount the drives in Windows/Linux and use partitioning programs that way
Sounds like the next challenge to mount it to my MAC, but thank you! I will google a bit when the CM4 is delivered.
https://yellow.home-assistant.io/guides/reinstall-os/
Option 2 is the way to access the storage via the USB-C port. Before flashing the software, you should be able to remove the NVMe partitions
Just one other question, i know not the right thread here. My /config path has somehow renamed is now homeassistant/ - Can you say if that is part of an update or did i change that with a setting?
It's been part of an update. Goal is to improve the backups of add-on data.
All right, thank you.
It might also work with the ha network
Hello, I'm currently setting up my yellow. I have the CM4 with eMMC and I have an SSD. Should I force install HA onto SSD instead of eMMC?
Hi. I'm trying to install a new compute module on my yellow. Both are lite versions, so it should just be to switch them out.. but no joy. The orange led is not glowing at all. Any tips ?
Check if the CM4 is fully seated in its sockets. Depending on the firmware version of the CM4, a firmware update might be needed.
https://yellow.home-assistant.io/guides/flash-cm4-firmware/
lol i lied.. it was a a eMMC i had.. and now i have a lite.. hehe sorry
In that case you need a downloaded backup and have to reinstall HAOS to the NVMe (might need a wipe of the partitions before, if you want to use the installer USB stick).
noted, i will try 🙂 thanks
Well Frick.. I got it online, but the backup file it did not like.... So I need to install it all again 🤣 fml
What do you mean with "did not like"?
I did not double check if it worked. So I think I cut the power to early hehe my bad
Restoring backups can take a lot of time
Anyone seen an issue where the serial-over-USB connection to their Yellow ceases to function? Mine was working fine at least a few weeks ago, but now when I connect to it, there's no data passed back or forth. The TTY device still shows up in both Linux and Windows, but I can't seem to get any data over it. I'm wondering if there's something borked with the serial config on the HA-side of the connection. SSH connections and the terminal plugin both still work, just not the serial-over-USB connection.
Hi @languid pulsar I also get an issue a couple of day ago on my Zwave usb stick on the yellow HW. IT's stops to worked. I cannot see anymore the stick on the "all hardware" list. So I expect an issue with the stick. I purschase a new one and it behave the same ! Could that be linked ?
Do you mean a black screen after connecting to the serial console or does it refuse to connect?
If it is the first one, press Enter a few times or send a Break before pressing Enter.
It connects (in so much as a serial terminal ever does), but it passes no data back or forth. Hitting enter or rebooting still doesn't generate any output.
Possibly? I've been having some issues with the built in ZigBee as well (#zigbee-archived message) which I believe is also a serial connection. I wonder if there's some issue with the serial bus. That said I also have a Z-wave radio attached to the gpio port, and it still seems to be working fine, so it's not like all the serial devices are broken.
I think for sure i need a uncompressed backup including the database because if i open the service call in the developer tools excluding the db is checked as standard?
Yeah, it is checked but the option is disabled. I unchecked that nevertheless and disabled the compression when using the service hassio.backup_full
Hi, I continue to get the following error when installing, I have tried reinstalling the OS several times and this error continues to occur, what can I do?
--- INFO (MainThread) [supervisor.docker.interface] Updating image ghcr.io/home-assistant/yellow-homeassistant:landingpage to ghcr.io/home-assistant/yellow-homeassistant:2023.12.4
--- INFO (MainThread) [supervisor.docker.interface] Downloading docker image ghcr.io/home-assistant/yellow-homeassistant with tag 2023.12.4.
--- ERROR (MainThread) [supervisor.docker.interface] Can't install ghcr.io/home-assistant/yellow-homeassistant:2023.12.4: 404 Client Error for http+docker://localhost/v1.43/images/ghcr.io/home-assistant/yellow-homeassistant:2023.12.4/json: Not Found ("No such image: ghcr.io/home-assistant/yellow-homeassistant:2023.12.4")
--- WARNING (MainThread) [supervisor.homeassistant.core] Error on Home Assistant installation. Retry in 30sec
You can try to ping google.com from the serial console to see if there is an issue with the internet connection or DNS
Hi, I have added an NVMe drive to my yellow, and now i would like my backups to go to the eMMC on my CM module, but for some reason it does not show up in the selector when trying to change location for backups?
If the data partition is moved to the NVMe, the eMMC is used for the OS only. You don't have access to it to store backups there.
@brave trout I converted your message into a file since it's above 15 lines :+1:
You can try to install HAOS directly instead of using the Installer.
Follow Option 2 of this guide, except the flashing part:
https://yellow.home-assistant.io/guides/reinstall-os/
Instead of using the Raspberry Pi Imager, use Balena Etcher for example and flash the following HAOS image to the Yellow:
https://github.com/home-assistant/operating-system/releases/download/11.4/haos_yellow-11.4.img.xz
Hi! I bought Yellow Standard, put my ssd in it, run through installation process but now I can not figure out how to migrate OS to that SSD. There are no nvme devices listed under "all hardware". Is it SSD issue or Yellow issue?
SSD is confirmed to be working outside Yellow
What specific SSD are you using?
Nothing from the "Not working" list on FAQ but will provide exact name in a minute
XPG SX6000PNP 512G
And is isn't shown as possible migration target here:
https://my.home-assistant.io/redirect/storage/
Hm, I can't send screenshot but it saying:
Disk metrics
Used space 31.6%
eMMC Lifetime Used 10%
Move Data Disk
Network Storage
No connected network storage
Add network storage
And if I press "Move data disk":
No suitable storage found
There is no suitable external device found. The storage capacity of the external data disk must be larger than the storage capacity of the existing disk
If you connect to the serial console and run lsblk | grep nvme , does it show anything?
You can try to mount it again. But it might be incompatible.
Yeah, I remounted it like 5 times already... Oh well that one was just sitting on the shelf and I was hoping I can make it useful again. Any recommendations for one that should be compatible for sure?
The Samsung 970 Evo Plus is a safe call.
There are also cheaper ones. But that's what I'm using too.
What size?
As much as you like. I'm using the 1TB one, but that's likely overkill for most usecases
Ok, good to know that 1TB works, I do like overkill 🤣 Ok, thank you mate, I'll source one like that and hopefully it would solve my issue
It is simillar to my 16% CPU, 14% RAM, 21% GPU on my PC 😄
Yeah, 1TB sounds big enough and I can see it on sale so I will go with it, and if it would help - would submit my XPG for incompatible list in FAQ
So this process flashes it directly to the yellow? I have a CM4 lite so need it to go to the NVMe, as there is no eMMC. (i.e. Balena Etcher does not see any storage.)
This should be possible directly after i connected the board to the computer or first when i run rpiboot?
Hi all, I created a feature request to monitor NVMe drives in a Yellow. If you think this can be useful, please vote for the feature.
https://community.home-assistant.io/t/monitoring-nvme-drive-health-in-a-home-assistant-yellow/673194
@sacred bloom @brave trout The command rpiboot -d should show the NVMe, if I remember correctly. If not, I'd go with a m.2 to USB adapter to flash it directly.
I tried this and have the same result. I did ./rpiboot -d mass-storage-gadget64, then flashed the SSD directly with the haos-yellow-11.4 image. I get the same error as above where the serverip is not set.
No working controllers found
PCIe BRCM: link down
BOOTP broadcast 1
DHCP client bound to address 192.168.19.29 (0 ms)
*** ERROR: `serverip' not set
Cannot autoload with TFTPGET
HAOS>
```What is it supposed to do?
I've got a samsung 970 500G on order and will try that. If it turns out to be the SSD, it is concerning that the system is that sensitive.
It looks like the CM4 firmware is trying to boot from network. A firmware update might fix it.
https://yellow.home-assistant.io/guides/flash-cm4-firmware/
If that doesn't fix the boot order, there are also options to change the boot order in the Raspberry Pi Imager.
I've changed the boot order (using rpiboot, not Raspberry Pi Imager) to BOOT_ORDER=0xf46, but it still seems to want to boot from network.
At the HAOS> prompt, I get this:
HAOS> nvme scan
HAOS> nvme info
HAOS>
it is not seeing the SSD. 😦
I've got the same one, but didn't install it for the moment 😂 does anyone know if I can use the yellow also as a cloud for storage? Thanks
It is not impossible (with a NextCloud add-on for example), but not recommended. The processing power of a CM4 is limited.
Which SSD model is this exactly?
HP SSD EX900 Plus M.2 512GB
PCIe Gen 3 x4, NVMe 1.3
Power Management
S.M.A.R.T. Supported
TRIM and ECC Supported
etc...
@brave trout I analyzed some NVMe which did not work with CM4, see https://github.com/home-assistant/operating-system/discussions/2235#discussion-4575642. Since the system ca not boot from it at all it is probably a Type 1, 2 or 3 failure.
At least some started working after a CM4 firmware upgrade, but it really depends on the issue at hand.
<@&330946878646517761> Spam-Alert User Psu798076
@brave trout I converted your message into a file since it's above 15 lines :+1:
What we can confirm though is that the Compute Module 5 will share the same dual connectors as the Compute Module 4
Retaining the dual connector should mean that boards designed around the CM4 should work with the CM5. The assumption being that the 55 x 40 MM form factor remains the same.
Where I am supposed to buy this Home Assistant Yellow from brand new in the UK? "Order now" only lists USA, Denmark, etc
I would look, if one of the EU or CH sellers ships to UK and if they offer VAT free shipping. You likely have to deal with customs and taxes yourself.
Hey, my new CM4 comes without new thermal pads but a thermal pad has come loose during assembly of the new CM4. Can I still use and assemble the new CM4?
You can still use those thermal pads. Or is it lost/defect?
It should be safe to use it without. The specs of the pads can be found here:
#yellow-archived message
The CM4 might start to throttle earlier without
Then i will order new one
@wary pivot Thank you for all your support. I have replaced the CM4 successfully without any issues except one defect thermal pad. I have choosen to Boot the new CM4 from USB first to flash the eMMC. Then i have deleted the partion from the NVMe with a USB adapter like recommend from you. That was super easy and super fast in total! Thank you also Home Assistant for the user friendly design of the Yellow which makes the upgrade really smart.
Apologies if this is slightly off topic, but I bought some Yellow components from Seeed Studio, and they're backordered (which isn't shocking), but I can't find any info about expected shipping times. Does anyone have any recent experience with this? Obviously, I'll contact them next, but I thought some here might know. Thanks!
Seeed usually shipped quite quickly
Thanks, that's good to know. It's only been a week, so I'll sit tight before asking, I think.
Maybe their customer support can give you an estimate
Also, does anyone suggest any other connectivity devices for a new Yellow other than SkyConnect? I realize it will likely depend on my desired applications, but anything common that those two systems lack out of the box? Feels like I'll have network connectivity, Bluetooth (with the CM), and Zigbee/Thread/Matter with SkyConnect, so I'll be in pretty good shape, but am I forgetting anything that will be frustrating if I try to connect a common device and realize I have to purchase and wait for shipping for a new dongle, do folks think?
Zigbee/Thread is on-board the Yellow. If you want to use both, the second radio Like the SkyConnect would be recommended. If you are using Z-Wave devices, you might think about a HAT for the GPIO pins or a USB dongle for that
Thanks once again - will check out the latter, as I am using Z-wave, yes
On-board BT has mixed results. Some are happy with that, some not. ESPHome BT proxies are recommended.
The Aeotec Z-Pi 7 works fine inside the case using the GPIO pins for me. Other brands should work as well.
Thanks for the recommendation - lots to research, and I appreciate it
Is the Home Assistant Yellow out of production?
I've got a very annoying issue, sometimes when my yellow reboots it gets stuck with the solid red and green light and I have to unplug and replug it to get it to continue. Anyone dealt with similar?
I only had an issue with a faulty NVMe disappearing until powercycling. You can observe the boot process via the USB-C serial console. Maybe you can catch an error message, explaining what is happening
Hey, so I tested with 1TB 970 Evo Plus and it works as SSD for Yellow. Who I should message to get my ADate XPG SX6000PNP on an incompatible list?
Either via mail to yellow@home-assistant.io or in the GitHub discussion https://github.com/home-assistant/operating-system/discussions/2235#discussion-4575642
Mail might be better. The last GitHub discussion post was from March
Hello. I have a yellow kit, with CM4 lite 8gm ram, and a nvme ssd. Tried to install HA the regular way, yellow LED remained on. So, I tried to flash directly to the nvme using rpiboot and raspberry pi image; the yellow (or rpi) didn't appear as a selectable storage in the raspberry image. Any idea how can I proceed?
Edit: tried another usb stick and plug into the lower usb port of the yellow, able to install the OS to the nvme. Now system is starting up. Hopefully everything can go on smoothly.
Alright, I'll try that next time it needs a reboot. For the moment a docker container running on my yellow controls my thermostat so I'll wait until after the cold weather is over XD
When I do that maybe I should also try reseating the NVMe drive. maybe blow some compressed air on the pins
Anyone here using a Jabra 410 (or other speakerphone) and Yellow for voice assist?
If your question isn't Yellow specific, you might have more success in #voice-assistants-archived
Thanks. But issue is getting it to pass through to local voice assist on the yellow. Have it successfully working on a satellite Pi4 but there's something odd about the yellow.
I'll give it some time and then ask over there. Don't want to crosspost and it's not urgent.
The Yellow's processing power is quite limited. Using assist fully local is not working too well.
Yeah. I'm starting to figure that out. I sent from a VM hassio that I had been running for a few years. Goal was to be able to turn my desktop off from time to time.
I think I'm going to build a unraid server and run HA from there.
If you are subscribed to the HA Cloud, their TTS and STT services are good
Oh apologies. I'm not running local assist. I meant "assist microphone"
I do use nabu voice services
I'm just trying to get the speaker/mic to pass thru to them
Ah okay. The Jabra is connected via USB? The audio port on the Yellow is speaker only
Yes. And they are recognized by Wyoming/mic assist. I can select it in the config.
But it's not waking up on voice.
You have the openwakeword add-on running?
I tested the Jabra on a satellite Pi4 and it's running great there. "also using openwakeword and nabu
Yes openwakeword is running and works when input from a satellite pi
I think, the people in #voice-assistants-archived might indeed be better in helping with that problem.
Thanks for engaging with me. Good to have a sanity check to see if it was something simple.
Jorg is so helpful. Thank you
Hi all, I am new to Home Assistant and just getting started. I have a question regarding the CM4 module for the Pi. I am ordering the POE Yellow and want to know if I will run into any issues with a wireless CM4 module I have or if I need a non-wireless one to ensure proper functionality as designed. Any advise or input would be greatly appreciated. Thanks.
Really, it's just another network inferface. You could just disable the wifi once it comes up and use only the wired port for your connectivity.
Thank you. I really appreciate that!
Officially the wireless versions are not covered by regulatory approbal. But functionwise it doesn't hurt - you still shouldn't use WiFi
Just getting around to setting up my Yellow (PoE). I have a CM4 Lite, and I cannot get anything to boot. Just a solid red LED - no console output.. any ideas anyone?
Should I expect console output - regardless of any other factor (?)
Check if the CM4 is correctly seated in its sockets. It should have made a satisfiying clicking sound.
What NVMe are you using?
Hi man - yes definitely satisfying click!!
NVMe was bought with the Yellow. But even with the NVMe not inserted - I should at least get some console output !?
There should be a console output. You might need a driver though.
Hmm... I am macOS - what driver might I need 🤔
If you don't see /dev/cu.SLAB_USBtoUART and /dev/cu.usbserial-110 (or /dev/cu.usbserial-210), you might need this driver:
https://www.silabs.com/developers/usb-to-uart-bridge-vcp-drivers?tab=downloads
I do not see a SLAB_USBtoUART
Trying this driver now..
I should be seeing console output pretty quickly after applying power to the Yellow?
I see /dev/cu.SLAB_USBtoUART now
You have to press Enter sometimes
I get nothing.. red LED no other signs of life
Even without anything installed (blank SSD)....? I should see something....
Any special USB-C cable required 🤔
Do you have an older hardware revision with a jumper at USB-C Rcvry (besides the m.2 slot) instead of a small button?
Yeah I have the older revision...
Is the jumper open or closed?
Open
Is the jumper below the USB-C port in position UART or USB?
Yes, the other position would enable access to the CM4
I have a CM4 2GB Lite.. but I think it should still work, right (?)
Where "work" means I see some console output as it "tries" to boot 🙂
Should I see any console output without CM4 inserted?
Trying to triage whether it is the Yellow or CM4 at fault here..
The console needs the CM4
Have you tried the install procedure with the installed NVMe yet?
I have tried booting to a USB image.. or do you mean installed onto the SSD separate from the yellow?
Is there any point trying to image the SSD, if I can't get any console output from the device?
You can try it. It wouldn't hurt. If you flash directly to the NVMe, you can also use the HAOS image for Yellow (with Balena Etcher for example) instead of the installer image from the Raspberry Pi Imager.
https://github.com/home-assistant/operating-system/releases/download/11.4/haos_yellow-11.4.img.xz
I need USB hardware to host the NVMe?
I can't get anything to boot on this device though (?)
Presumably I'd need to use different hardware to write the NVMe
The easiest way would be a USB to m.2 adapter
Yeah, don't have one sadly
If I have no NVMe installed in the Yellow I would expect to see errors on the console...
(I have tried - silence)
What USB image have you used? The Installer from the Raspberry Pi Imager?
Yeah..
But I'm not even getting to the point where it's even looking at the USB yet...
One strange thing: macOS built in screen thinks the serial resource is busy
Brew's screen doesn't report this error
(just a blank screen though)
You can try to use the Yellow as a USB adapter. Basically this guide - option 2 (until step 7)
https://yellow.home-assistant.io/guides/reinstall-os/
With the command ./rpiboot -d mass-storage-gadget64 you should get access to the installed NVMe
OK - but even without the NVMe installed I should see output on the console, yes?
ie. that test removes the NVMe state from the equation..
I have no CM4 Lite to test that
Right i c
What NVMe has been sent to you? Brand and type?
WD 256GB
1sec
PC SN530 NVMe™
SDBPNPZ • 256 G
FW: 21118000
Trying rpiboot
❯ sudo ./rpiboot
RPIBOOT: build-date Jan 20 2024 version 20221215~105525 20a431c6
Waiting for BCM2835/6/7/2711/2712...
(nothing further)
Not really sure what I should do - buy another CM4, or assume the Yellow is just defective
Then again, the rpiboot should have worked without CM4?
No
Just to be sure - you have set the jumper below the USB-C to USB and closed the jumper besides the m.2. When powered, the red LED is lit and the green LED went on, when opening the jumper besides the m.2
No - the red light stays on, green never does.
The rest is accurate (closed / open of recovery jumper) is correct
So green LED means "USB SSD" mode?
In that specific case, yes
Worth buying another CM4 - or are you dubious this is the problem?
It just seems that nothing is working - there is not just one thing..
I would remove the CM4, inspect the contacts and plug it into the sockets again. Just to rule out a misseated CM4.
If that doesn't work, I would recommend to contact the support at yellow@home-assistant.io
already tried re-seating.. thx
I'm not 100% sure, if a NVMe problem (like an incompatibility) could "break" the serial console on a CM4 Lite. The support might know that
I would wait with buying a new CM4 until support has looked into it. They might have an idea, I have missed (totally possible - I'm at work right now 😅 )
No prob - appreciate your help!
I have a Yellow with a CM4 Lite and an NVMe. Also could not get it to boot. The solution was to get an M.2 enclosure and install HA on the drive directly from my PC.
Just FYI for posterity, I got a shipment notification from DHL saying that it’s on its way from Shenzhen and due to be delivered just under two weeks after I ordered it. Not bad at all!
Hopefully I can get it to boot …
<@&330946878646517761> Spam alert, user $in ^ (edit: not SnarkNerd - just do be safe)
Thank you!
I had a similar issue at first. Is your CM4s EEPROM updated? That’s what I needed to get mine working
Looking at your earlier comment about nothing happening, you’re not specifying the directory for the update file, that’s why it’s not working. Try this #yellow-archived message
Tl;dr try
sudo ./rpiboot -d recovery
Oh wait.. you’re not even getting past the point where it recognizes the CM4 itself. Have you tried a different carrier board? There’s cheap ones on Amazon
I don’t have another board. Others suggested imaging the M2 separately - do you think that might work? There is an EEPROM on a CM4 Lite that you can update?
I should be getting some signs of life out of the Yellow LED's if the CM4 is operable, regardless of anything else? (ie. not just constant red LED) is that right?
If this doesn't work, seating or broken CM4 are the most likely cases. There is essentially no electronics on the Yellow active for this to work (USB is essentially directly connected to the CM4). Also we do test this function in the factory.
When reseating, make sure to double check both sides of the connector for damage. Check alignment carefully. When pressing the CM4 on, make sure to take the PCB off the transparent polycarbonate base, so you can press against the PCB directly as shown in Step 3 https://yellow.home-assistant.io/poe/#installing-the-cm4
If you still can't make progress with rpiboot and connectors look clean/undamaged, I would say CM4 is more likely.
Thanks @shut cedar!
Yes the lite has eeprom. I only have CM4 Lites
Is there documentation somewhere for what voltage/wattage the POE supports?
I'm trying to mount an external USB hard drive on a Yellow. I plug the drive in, /dev/sda and /dev/sda1 appear (with appropriate messages in dmesg) but neither mount nor fdisk can open the drive. Any ideas?
So next question, it seems I might not be at the right command prompt to be able to do such stuff. I followed the instructions here: https://developers.home-assistant.io/docs/operating-system/debugging/ to enable ssh access to the actual OS, but port 22222 remains closed, even after a reboot.
PoE+ IEEE 802.3at-2009 Class 3 or 4
Up to 25.5W PD input at 48V DC, max 600 mA
Hi, I see recommendations to use USB sticks to use with the Yellow. So is the thread radio in Yellow considered unsupported? Or is there a possibility that it will be stable some time?
In Yellow the only way to use Thread is to enable multiprotocol. Which seems against recommendation
That is not correct. The radio inside the Yellow can either be used as Zigbee only or Thread only radio. The Multiprotocol add-on is a way to use the radio for both simultanously - but is still experimental and not recommended for production use.
Neither option is unsupported - the Multiprotocol option is just not considered stable yet.
If you want to use Zigbee AND Thread, I would recommend to add a second radio (like the SkyConnect), instead of the Multiprotocol stuff.
That's not what this alert is saying when I go to the option to remove multiprotocol
I can't attach an image
It will take a few minutes to install the Zigbee firmware and restoring a backup.```
Any pointers on how I can go back to Thread only radio?
Remove all ZHA/Thread integrations and the Multiprotocol add-on and use the OpenThread Border Router add-on instead. It will install the Thread only firmware.
Thanks, that did install new firmware. It's also asking me to install the Thread integration again. Should I do that?
Yes
FWIW, we have instructions for how to go from Multiprotocol to Thread, the recommended path is disabling first and then go through this: https://yellow.home-assistant.io/procedures/enable-thread/
Thanks. For some reason I always ended up on the multiprotocol doc page and that only mentioned going back to zigbee. Should have put more effort into a search
Hello everyone, I have a question. Is there any update on the final product to solve the problem of CORAL not working? I understand that the CM4 board was the issue, but do you have any idea if the Raspberry Pi boards will receive an update to address this in the near future (within 1-2 years)? Thank you.
A hardware fix for the m.2 Coral of the CM4s is unlikely. The USB version does not work reliably as well.
Thanks for adding this. It wasn't there when I was setting up my Yellow a couple of weeks ago.
Hi I am new to the scene and trying to combine a few of mine devices from different plaform to homekit and home assistant. I have issues moving smartlife IR blaster to homekit and after trying to find out more was direct to homebridge. I currently have a home assistant yellow box how do i install homebridge is it via the Raspberry PI option?
On Raspberry Pi 5 the Coral M.2 works, so maybe with a CM5 🤔 😅 But if that will be compatible with Yellow remains to be seen 🥶 https://www.jeffgeerling.com/blog/2023/pcie-coral-tpu-finally-works-on-raspberry-pi-5
Quite some substantial changes, if the rumours are correct. USB3 instead of MIPI lanes for example.
First thought about that (not even thinking about pinout compatibility): more potential interference
Interestingly, despite all the warnings, the Silicon Labs Multiprotocol v2.4.3 is stable and working on my HA Yellow. Zigbee and Thread both working fine. Maybe because I also have a couple of Google Nest border routers on my Thread network?
hi all, FWIW i tried installing HA onto a new Yellow POE kit this weekend and was getting a solid yellow light no matter what i did. it never got an IP address/connected to the internet. it turned out to be the m2 SSD drive i got from Seeed, i removed that and everything installed fine.
for anyone having this issue i would definitely try removing the SSD, and it seems like there's a particular issue with these SEED ones
For what it is worth, I got a Samsung 970 EVO Plus SSD today and wrote the haos 11.4 image to it in an enclosure with BalenaEtcher. Moved it to the yellow and it booted up with no issues. The HP EX900 Plus SSD is going back.
Thanks for your help @Jorg and @shut cedar , I really appreciate it. I am running now with a Samsung 970 Plus SSD.
Had my Yellow die overnight. No lights or anything. Tried reseating CM4, no dice - still no lights. Any ideas for testing I can do?
@fathom breach no red light? Tried a different power adapter?
Haven't tried a different power adapter. But no lights whatsoever
Tested w/ a multimeter and seems fine; getting 12V as expected
Well this unbelievable - the NVME screw on my Yellow snapped in two and now my drive won't seat securely. How can I obtain a replacement screw? Is there a code number to the screw I can lookup and order somewhere?
At least the standoffs are custom made. Did the screw or the standoff break?
Hi Jorg it is the post inside the nut that broke. I was so careful not to over tighten, but it just snapped and I heard a piece hit the table, leaving another piece stuck inside.
If the hole is of M2 size, would a standard nvme screw work?
the post separated from the head?
I don't know what the right terms are
It's an M2, 3mm, 4mm head
Correct. There is a post screwed inside the head (part with the groove) and it just snapped off.
I managed to get the rest of the post out. Now I have the grooved head with no post :/
M2 3mm sounds right for the screw in part.
the stand off is unfortunately custom, but the screws should be something you can find
I think Ace Hardware still reigns the king of the smallest screws. I will just get a bunch of 3 mm width different lengths and trial n error it until I find the right one.
Simple question: I just ordered my yellow today. I’ve got a couple of NVMe drives sitting around, and I plan on using one. Do I need to reformat to EXT4 before trying to use it, or will the setup process automatically format the drive (and nuke any data in the process)?
If there isn't a HA installation or data partition on it, the Yellow should nuke it. But I removed all partitions nevertheless - without filesystem afterwards.
Yeah, these are ones I pulled from laptops or something when I first got them. They have windows or something on them but they were never used by me. Since it’ll nuke it I’m good to go. Thanks for the quick response 🙇
As someone who was planning to set up a Yellow with M2 from Seeed this weekend, thanks for sharing this (and yikes)! Did they accept a return and give you a refund?
no I haven’t gotten around to starting the return yet, fingers crossed there’s no issue
it had to have been the m2 drive, it was the only change and install was perfect as soon as I removed it and reset the board
Heya, so I got the HA yellow kit (the kit with the CM4 preinstalled, but the m.2 ssd separate in the package) from raspberrypi.dk and when I initially connected it, it didn't show up in my network.
So I assumed the OS is not on there yet, so after that I followed the guide to install the raspberry pi os on it, but I have no idea how long to wait. Does anyone have any pointers for that? The yellow led has been steadily going for a bit, but how long do I approximately have to wait until I know it is or is not working out?
It's been going for 20-25 mins now I think
What SSD exactly did you buy?
The default they offered on the site. A 500gb kioxia
I meant the default on that site: https://raspberrypi.dk/en/product/home-assistant-yellow/
All right. I'll give that a try.
It's (presumably) starting up.
If it starts and onboarding is completed, shut it down, insert the SSD again and look if it starts and shows the SSD as target to move data partion to.
If not, I assume the SSD is incompatible - there has been another user with problems using that one
I've still got a 2tb samsung 980 pro laying around, but that seems very excessive for this use case.
What shows that it has started/onboarding is completed?
The Samsung 980 Pro is confirmed incompatible
Ah, I see yeh
When you get the login screen after the setup using http://homeassistant.local:8123
it seems to be doing the same heartbeat-like yellow led thing, with a very short green pulses and continuous red
