HA Restart hangs if wyoming satellite not available
jp442 opened this issue · 5 comments
I installed Wyoming satellite as per the tutorial https://github.com/rhasspy/wyoming-satellite/blob/master/docs/tutorial_2mic.md
Now, if I physically turn off the Raspberry Pi and then restart HA, it hangs at the point where it is trying to load Wyoming saying "Starting Wyoming Protocol, not everything will be available until it is finished"
I try to dismiss this, but I can't do anything in HA until I turn the Raspberry Pi back on. The only other option is to disable the device within the Wyoming Integration before I do the restart or reboot.
I’ll look at it.
I noticed something similar as the integration will hang with a red failed but just the device.
The others entities (piper, etc) loads normally and HA works normally also.
Does HA freezes completely? This is very unusual.
I can navigate to some of the pages in the GUI like Settings, Developer Tools, Studio Code Server, but File Editor, Zigbee2MQTT, and the Overview page won't load (I didn't try them all). Also, none of my automations work. The only fix is to disable the device in Wyoming Integration and then reboot HA. Alternatively, I can reconnect the Satellite and wait for the Wyoming integration to heal itself by re-establishing communications to the device. It appears that if I wait long enough (several minutes) it does eventually complete the startup and everything seems to work again (except for the satellite of course).
Further info on testing the satellite. If the satellite is not operational when HA restarts or reboots, it takes 5 minutes for the load of the Wyoming protocol to timeout before HA is fully operational. Also, if HA is running with a failed satellite device because the satellite wasn't powered on when HA restarted, turning on the satellite after the fact doesn't work. Wyoming still shows a failed device, no matter how long you wait. The only way to reconnect the satellite is to disable and re-enable it in the integration.
Hopefully someone will be able to look at this and fix it.
Is anyone else noticing this issue?
Is anyone working on a fix to the problem?
Am I the only one with this issue?