Raspberry Pi Reboot error - V #129714
Comments
|
same problem |
|
Same issue for me on the pi5, tried with 2 different nvme boards:
The issue only occurs after a reboot, if I disconnect/reconnect power it boots up fine. |
|
Same problem for me with a clean install of HA with raspberry pi imager and use of sd card only. Sometimes it works, sometimes the boot loop hits. |
|
Same problem here |
|
One thing to add is, that the issue only occurs if I restart the system via webgui. If I reboot it via „login“ shell (port 22222) the raspberry boot up normaly. |
|
Same issue with rpi 5 and Argon one v3 nvme case |
|
I've done 50x ssh root@IP -p 22222 "reboot" and 50x ssh root@IP -p 22222 "ha host reboot" -> every reboot was succesful and it booted up normally. Only the system reboot functionality in the web gui causes the boot problem sometimes likely every third time |
|
Same issue with rpi 5 and Argon neo 5 nvme case You need to manually power off and turn it back on for a normal boot |
|
[Hey] there, Attempt 1: Raspberry Pi 5 (8GB) Attempt 2: Raspberry Pi 5 (8GB) This issue has been happening with both setups for the past 1-2 months. ( |
The automatic reboot after updates seames to work too, only manuell reboot in Web ui has problems |
I first noticed the issue after a nightly reboot from an automation (that uses the hassio.host_reboot action) could be that the Web ui uses that same method of course (idk). |
|
Same problem here any update ? |
I've tried running the "ha host reboot" command in the SSH add on and experienced the same issue. Have you tried this as well in the SSH add on? |
Nope, I‘ve done this directly on the host. Maybe the problem occurs only if the reboot is triggered in the docker container |
|
I share the issue as well, every 2-3 reboots initiated from the UI lead to this 
|
|
I have this problem too. Case Argon Neo 5 NVMe. Connected to USB 2.0 SkyConnect, Bluetooth (CSR8510 A10 0a12:0001) and RTC accumulator. Every 2-3 rebooting Pi5 not see 'nmve'. I try set only BOOT_UART=1
WAKE_ON_GPIO=0
POWER_OFF_ON_HALT=1
BOOT_ORDER=0xf416
PCIE_PROBE=1boot config: ...
dtparam=nvme
dtparam=pciex1_gen=3
...
UPDATE |
|
i have seen on both pi5 with the ssd boot problem the status of boot slot b is "bad"
|
|
|
I have the same result as @user86000. |
|
So, I'm tired, install the native operating system with GUI and deployed proxmox in it, disabled all updates. Then installed the home assistant virtual machine (haos-aarch64.qcow2), restore backup and all the problems disappeared. Now there is native access to ssh, vnc, eeprom and the config.txt file. A million reboots and everything is fine! Now I'm happy, on NVMe work is very fast!
|
|
Exactly the same issue for me. Also tried to flash a different sd card with Raspbian, did a firmware update by doing following: Still no luck. |
|
Been having the same issue for a long time. rpi5, Argon NVMe case. Stuck most of the time on reboot in bootloader trying to boot from NVMe, gets resolved by cycling the power. I updated the bootloader (booted from SD card to be able to do that) to the latest, just to see if it changes things. It does not. |
if install native RPi OS with gui and install script from Argon manual then trouble is gone. Million reboot and nobody problem! |
This comment has been minimized.
This comment has been minimized.
|
Same here with boot from sd card, tried bootloader update but nothing changes, fails after 2/3 reboot from ui. |
Thank you for the suggestion 👍 I was not 100% sure I ran those scripts on this raspberry, i.e. these: https://download.argon40.com/argon-eeprom.sh So once again, I booted my HA-rpi5 into Raspberry Pi OS and ran those. (And once again, updated the bootloader to the latest with a nice bootmenu.) Then retried booting from HA UI. First time fine. Second time stuck. Oh well. So unfortunately that did not solve the problem. As a workaround, I made a hardware watchdog out of a Shelly Plug, that automatically cycles the power of the rpi if it gets stuck on boot 🤓 Gist here https://gist.github.com/tpln/0f72b2cb1c70b1db6e885ade166fec8f. |
Initially, I did exactly the same and it didn't help me either. For the experiment, I installed a native operating system with a GUI to NVMe and executed Argon scripts. After that, I deleted the flash card. And tried to reboot more than 10 times. And each time there was a successful boot, 10 out of 10! Only after this experiment I switched to Proxmox. Now every fine! |
|
I have the same issue. Case Argon Neo 5 NVMe. |
|
Same issue here without SSD. Just plain RasPi 5 with micro-SD card (no SSD at all) |
|
It appears that the issue is not related to the SSD; it may be associated with the Raspberry Pi 5? as I have encountered no problems while operating on a Raspberry Pi 4. |
|
Hmm, sure enough, this morning went to test a reboot and it hung up again. This time was a bit different. Attached a display but nothing came up. In the past, I saw the bootloader cycling through boot options and display the nvme error before the hard reset after max_restarts kicked in. This time it was just a blank screen. I didn't pay attention before, so this may or may not have been the case before, but the rpi5 was also flashing the green led 8x quickly, which according to the rpi docs suggests an SDRAM error. Reddit users claim that to be a h/w issue, but at least in my case, a hard power cycle by removing the usb-c usually resolves the boot problem. I decided to boot back into the OEM Raspberry PI OS lite and see if any of jonyrh's tweaks would help. But get this, after making the changes, even Raspberry PI OS lite got stuck booting after a "sudo reboot now", with again nothing on the display output and only the 8 green led blink pattern repeating. Tested that a few times and was able to reproduce it (inconsistently), so that would seem to rule out that this is to specific Home Assistant OS and maybe this is an upstream issue. A bit confused now. I'm not sure what was different with my bench setup when I was testing the before fix and it survived so many reboots. Went back to my bench setup to muck around a bit. The max_restart setting is definitely helping, as a few times I can see the bootloader attempt to start from the NVME and fail, and after a few retries, the hard reset resolves the problem. But there are other times when during a reboot cycle, it seems the bootloader itself might be crashing, the one where I get no display and only the 8 green led blinking pattern. When that happens, the only resort seems to be a hard power cycle by unplugging. And a few times, even that was finicky, and I had to unplug everything from the RPI5 before the bootloader would come back up. I had tried the use latest bootloader firmware option in raspi-config. Now I'm wondering if maybe switching to the latest made this worse, or if its the settings. Will try a few variations today to see if I can get it any more stable across reboots. |
|
No luck, giving up for now. Tried both the latest and stock bootloader from raspi-config but both seems to hang intermittently. For now, just left the stock firmware and added only the MAX_RESTARTS flag, since that seems to address the nvme issue some of the time when the bootloader hasn't crashed. May just go with tpin suggestions and put hardware watchdog on it. Or just return it all while I'm still in the return window and try some different hardware altogether. |
Also if rebooting with an automation. I do a nightly reboot to keep everything nice and fresh but every morning my Pi5/nvme SSD hangs and requires me to unplug the Pi. |
|
Hey guys of Home Assistant ... this is a real nuisance ... can one of you chime in please? |
|
Same problem |
|
Issue resolved for me, when i removed the Sonoff USB ZBDongle-E and started using an IP based SMLight zigbee device. |
|
I don't use a Sonoff USB, but I do use a Zooz 800 USB for my z-wave devices. Are the others having this problem also using some similar USB devices? |
I have this problem on both rpi5 ha instalkations, one without any usb device. The other one had an USB skyconnect device but with disconnected the same boot loop too. |
|
I have the same problem with HASS on sd card and usb SONOFF Zigbee Dongle-P but it seems like the usb have nothing to do with the boot. I tried removing the dongle but after 2/3 reboots from web ui the problem is still here. |
|
yeah, having the same boot problem ... I use a SkyConnect USB Zigbee/Thread Dongle. Will try without and see if fixed. |
|
Hi , anyone have some solution ? I'm not an expert but i have the same problem in both my raspberry pi5 home assistant OS installations. Thank you |
|
Can anyone fix this? We rly need to get attention on this |
|
Same issue here on a RPI5 with homeassistent os on a SD card. |
|
Confirmation cases:
Behavior: Issuing a host reboot from Home Assistant's UI can cause the Raspberry Pi 5 to get stuck in the bootloader, while trying to find a valid partition to boot from. The bootloader fails to recognize valid boot partitions and continues to loop through all boot media options, SD card, NVMe, and USB flash drive. This, strangely, only occurs about 50% of the time. Workaround: My current workaround is to use matter enabled Eve Energy Smart Plug to be able to kill power to the Rpi, when a remote reboot need to be done. I added the smart plug to Google Home, using Nest Wi-Fi hardware's builtin thread border routers. |
|
Also can confirm issue is ongoing. I had two problems, One it would not reboot as others have said, then the second was when it did boot it would get a ping bot not boot any further. I got a new GeeekPi Starter Kit for Raspberry Pi 5 8GB, with N04 M.2 NVMe SSD PCIe Peripheral and a 1tb m2 nvme. Built it up and restored my backup. Power ups are now fine, boots 100% of the time. I think my last SSD was faulty. But if I reboot the pi from the reboot option about 50% it will not boot getting as others have shown on their screen shots. Does not seem to properly detect the nvme. Glad I am not alone on this. |
|
i have a same issue RPI5 8gb |
|
Same Home Assistant issue, only on "Reboot System" does it hang. Fresh instal of HA. I've tried 2 different stock (but fully updated) Pi 5's, two different NVME boards (official and Pimo Roni), and two different NVMEs (Sabrent rocket and Sk hynix). Testing was with the official power adapter, no USB's plugged in, no SD card. Changing boot order or leaving stock had no change. |
|
Add me up to this list. The reboot worked in the beginning (I bought the PI and the HAT at launch), stopped to reboot several months ago. Only way is to unplug power plug, like everyone else has already said. RPi5 |
|
Add me to the list too. Until I discovered this thread I was working under the assumption I either had an SSD compatibility issue or it was a HASSOS boot problem. My latest idea was that it had something to do with the A/B boot scheme being used by HASSOS, but after reading this thread I'm not so sure. If I burn the latest 64-bit Raspbian image to the SSD, I can reboot all day long and I've never seen it fail. But rebooting from HASS, it hangs in the same boot loop as you folks are seeing. Yanking power and booting up again is always successful. I've tried stripped HASS down to the bones, disabling almost all the integrations, but that hasn't made any difference. I've been debugging this with two identical setups, both display the same behavior: sudo rpi-eeprom-update sudo rpi-eeprom-config -e I've also tried the following, but didn't see any difference: I've got some more parts coming in the mail to debug with. In the mean time, I'm going to try some of the ideas mentioned in this thread. Will let you know if I find anything. Just felt like I needed to chime when I found this. I don't feel quite as hopeless now that I know I'm not alone. FYI. When I started looking into the HASSOS boot status, this is what I always see. But maybe I haven't looked after a "successful" reboot: |
|
Forgot to add... Both Pis seem to reboot fine when running HASS from an SD card |
|
Not more to add other than me too. |
|
Form me sometimes it is stuck in the bootloop forever, sometimes it takes a few minutes and then loads.
|
|
Same problem with reboots hanging for me too (Been happening for months, I never got around to raising an issue, sorry!):- |
|
I guess this is my solution for now - using MAX_RESTARTS suggestion from tejesh83 above, and whittled down the USB wait time to 2 seconds. Adds 5 or 10 seconds to the restart time to have it fail and cold boot, but it seems reliable enough so far. Doesn't seem to matter if the Skyconnect dongle is attached or not. Gets me up and running for now... sudo rpi-eeprom-config FWIW, I don't really need a USB boot, but I've left it in the boot order just in case. Changing it to the following could shave a couple of seconds off the fail/restart time: [all] |
|
In case it helps someone debug this, I'm seeing memory allocation failures like the following in dmesg that aren't present when I boot the same hardware with the Raspbian OS. I don't have enough knowledge about the kernel startup process to understand if this is to be expected or not, and not immediately finding any related threads out on the interwebs that talk specifically about HASS.: [ 0.802928] cma: cma_alloc: linux,cma: alloc failed, req-size: 4 pages, ret: -12 |
I found some guidance here and here. Adding the following to /mnt/boot/config.txt resolved this issue (for now). Not sure if this is related in any way to the restart failures. Just trying to get the logs as clean as I can at this point: #Increase the CMA allocation |
that works for me! Not the first time around, as the settings haven't been applied to the running system yet, but the next reboot via the UI worked fine. |
I tried new install haos 14.1 with param |
|
Earlier I said;
Unfortunatly it does not, issues still appears, from time to time.
Had this setting from the start, no solution for the reboot problem |
|
I am new to Raspberry Pi, usually running my Home Assistant in a VM. Now, I wanted to explore RPI so I bought one. I flashed the microSD Card with Raspberry Pi Imager, taking the Home assistant image, from a Windows PC. Since this system is not in "production" yet, I issued a shutdown command from the I don't have a micro-HDMI cable at the moment to see if my message matches yours, but I'm a bit confused since my network interface responds. I remember this behavior occured back in December the first time I tried it, but I thought it was just a weird issue and I didn't configure anything then, so I just re-flashed it. Now, it's a bit more annoying since I did put a little bit of work in it. Maybe my problem is different, but it does seem to match the behavior of you guys. |
and others
The problem
When rebooting Home Assistant on the raspberry pi 5 with an NVMe HAT the device fails to detect the NVMe on startup.
This issue only occurs when rebooting, when powering off the device and then starting up the device by pressing the power button it starts up as normal.
For details on troubleshooting please see the home assistant community link below:
https://community.home-assistant.io/t/reboot-error-nvme-not-detected/781485
What version of Home Assistant Core has the issue?
core-2024.10.4
What was the last working version of Home Assistant Core?
No response
What type of installation are you running?
Home Assistant OS
Integration causing the issue
No response
Link to integration documentation on our website
No response
Diagnostics information
home-assistant_2024-11-03T08-17-41.453Z.log
Example YAML snippet
No response
Anything in the logs that might be useful for us?
No response
Additional information
No response
The text was updated successfully, but these errors were encountered: