Complete meltdown with Z-Wave, entire HA toast, can't boot

Been using HA for years. Running on an Pi 3B+ with a Aeotec Z-Stick Gen5. For the last 6 months, it’s become unstable requiring power cycle every 3-5 days. Z-Wave crashes out and then eventually the whole thing inaccessible. Comes back to life upon reboot. Rinse and repeat.

After updating to the most recent version yesterday, I had to power cycle 4 times to get it to come back to life, just long enough to grab the backup off the card before it crapped out again. So I figured the SD Card must be toast.

Got a new high endurance SD card. Imaged with the Pi imager for Home Assistant. Restored the backup. All was well again for about 12hrs. Overnight it just died with all connectivity and functionality lost. No UI, no samba shares, nothing. No amount of power cycling restored normal operation.

Imaged again and booted up fine. Restored backup, dead. Did this 3 times. Now I figure something in the backup must be really bad.

  • Reimaged and did NOT restore the backup. Went through initial new setup steps.
  • As soon as I tried to configure the Z-Wave, it died. Got as far as automatically making the security keys. Then it stopped, no UI, no nothing.
  • Power cycled and booted back into the UI. As soon as I went into Z-Wave config, it crapped out.
  • Booted with the Z-Stick removed and it was stable.
  • Booted with Z-Stick inserted, never even got a UI. Dead.
  • Factory reset Z-Stick (FML). Seemed stable, but hasn’t added any devices yet.
  • Removed Zwave JS and installed Zwave JS UI.
  • Won’t even boot to UI again.

WHY!!!

All SD cards go bad.
Make sure you have an industrial grade class 10 one from a reputable source. At some point that will go bad also. They are just not designed with rotating bits and failure correction in mind.

Get an SSD with a USB and have the USB connected to an external hum with it’s own power supply.

I hope you had the original ones back-upped as well. If you generate new security keys, you can no longer access all z-wave devices that were connected to the stick. Maybe it is freaking out over that?

Z-waveJS-UI has a config setting to backup both the settings and the nvm in the stick. If the HA backup restored half, than even a cleared stick might not ware well with z-waveJS-UI still remembering all devices?

But then again, without restored backup and cleared stick I would not expect this. Unless maybe the z-wave devices are still hammering the stick with data it cannot decrypt? I’m also just guessing here.

I already knew I may have to redo the entire Z-Wave network. That’s ok. I factory reset the Z-Stick hoping for fresh start with these problems.

The problem right now is any HA comms with the Z-Stick causes the whole thing to die. With ZwaveJS, everything is inoperable. With ZwaveJS UI, the whole system reboots and all settings fail.

Sounds like the z-wave stick itself could be bad.

This is driving me nuts. It started to cooperate. Redid all my z-wave devices in ZwaveJS-UI which is really nice. It survived a few reboots and power cycles. Disabling soft reboot seemed to make it work ok. Plugging into my windows PC to save a backup works fine. Zensys tools connects just fine.

Now today, back to being a dumpster fire in Home Assistant.

  • Boot with stick plugged in, then manually start ZwaveJS-UI add-on. HA completely explodes and reboots as soon as I start ZwaveJS-UI. No log entries since the last time it worked properly

  • Boot without the stick plugged in. Manually start ZwaveJS-UI. Everything remains stable. Of course this is useless without the stick. No log entries since the last time it worked properly.

  • Boot without the stick plugged in. Manually start ZwaveJS-UI. Everything remains stable. Plug in Z-Stick. HA crashes and reboots.

This is from the hardware page in HA:

ttyACM0
/dev/serial/by-id/usb-0658_0200-if00
Subsystem:
tty
Device path:
/dev/ttyACM0
ID:
/dev/serial/by-id/usb-0658_0200-if00
Attributes:
DEVLINKS: >-
  /dev/serial/by-id/usb-0658_0200-if00
  /dev/serial/by-path/platform-3f980000.usb-usb-0:1.3:1.0
  /dev/serial/by-path/platform-3f980000.usb-usbv2-0:1.3:1.0
DEVNAME: /dev/ttyACM0
DEVPATH: /devices/platform/soc/3f980000.usb/usb1/1-1/1-1.3/1-1.3:1.0/tty/ttyACM0
ID_BUS: usb
ID_MODEL: '0200'
ID_MODEL_ENC: '0200'
ID_MODEL_ID: '0200'
ID_PATH: platform-3f980000.usb-usb-0:1.3:1.0
ID_PATH_TAG: platform-3f980000_usb-usb-0_1_3_1_0
ID_PATH_WITH_USB_REVISION: platform-3f980000.usb-usbv2-0:1.3:1.0
ID_REVISION: '0000'
ID_SERIAL: '0658_0200'
ID_TYPE: generic
ID_USB_DRIVER: cdc_acm
ID_USB_INTERFACES: ':020201:0a0000:'
ID_USB_INTERFACE_NUM: '00'
ID_USB_MODEL: '0200'
ID_USB_MODEL_ENC: '0200'
ID_USB_MODEL_ID: '0200'
ID_USB_REVISION: '0000'
ID_USB_SERIAL: '0658_0200'
ID_USB_TYPE: generic
ID_USB_VENDOR: '0658'
ID_USB_VENDOR_ENC: '0658'
ID_USB_VENDOR_ID: '0658'
ID_VENDOR: '0658'
ID_VENDOR_ENC: '0658'
ID_VENDOR_ID: '0658'
MAJOR: '166'
MINOR: '0'
SUBSYSTEM: tty
TAGS: ':systemd:'
USEC_INITIALIZED: '5779376'

Have same problem as you, pi3b+ ssd, tryed everything posible and impossible, updayted sticks firmware, tryed to switch to different usb driver (rolled back as it didn’t help) and powered up sticks (zwave and zigbee) from external supply.

So for a moment “stable” if i can say so, setup was to disabling controller soft reset in zwave js, disabling automatic controller recovery and enabeling safe mode. With this options it’s working.

Hovewer after a day my zigbee disappeared from HA…and i started all over again, bootloop, HA safemode, changing drivers disabling addons and so on…finaly and again got it back to work without zigbee stick and with those 3 disabled options mentioned before…

I don’t know what can cause instability but may be that will help you somehow.

If you’re running OS 13+ try downgrading to version 12.2. I have a number of Pi3B+ that are unstable on versions of Home Assistant OS > 12.2