#Thread Devices Keep Becoming Unavailable

1 messages · Page 1 of 1 (latest)

elfin belfry
#

I'm running Home Assistant using the OVA in VMware ESXi. All of a sudden my Matter devices exposed via Thread keep becoming unavailable until I restart the VM. The Thread network is all HomePods and AppleTVs as border routers and a latest gen wired AppleTV is the device used for credentials.

How can I track down the cause of this sudden issue and fix it for good?

#

Network stack is all UniFi switches and pfSense for the router. None of that has experienced a change lately

elfin belfry
#

I would really like some help with this. No one ever responded and this is happening again

ancient canyon
#

what do your logs in the matter server for HA say as well as OTBR?

#

I've found that sometimes the OTBR addon stopped, so I have watchdog turned on myself

elfin belfry
#

Most thread devices come back online but usually 1-3 of them get stuck unavailable for some reason. I also now have one that I cannot repair after removing it in an attempt to get it back.

ancient canyon
#

My experience with matter and OTBR is that its not great

#

and these issues youre running into are what I've ran into

elfin belfry
#

So I should avoid further Thread/Matter devices for now as this will continue?

ancient canyon
#

All I do is make sure that both addons have the watchdog checked and it's been working. I sometimes have to use the native app to wake devices up to get them talking on OTBR again

ancient canyon
elfin belfry
#

I did start setting up a ZWave network using some box (I forget the name) as my Hass is a VM in ESXi.
For now though I'd like to figure out how to get Thread/Matter working again and just try not to mess with it until I can replace those devices

ancient canyon
#

Start by watching the logs

#

and writing up issues against otbr and matter server when you find issues

#

there really isn't much that can be done to troubleshoot issues from what I've found

#

you're at the mercy of those addons working properly

#

which should just "work" without fiddling

#

I'm running my matter and otbr on a shitty Pi4, and it takes about 3 minutes for otbr to start after reboot as well as matter server

#

once it's up and both addons haven't crashed, it just works. But again, I've seen issues where both crash randomly. So keeping the watchdog turned on will ensure the addon restarts when it crashes

elfin belfry
#

This all started from the wifi AP being rebooted, despite all Thread devices being via Apple Home hubs (and the primary hub being wired Ethernet)

tribal osprey
elfin belfry
tribal osprey
#

Super interesting, I have nothing to add to the thread, but has the AP been restarted again?

elfin belfry
#

Yeah the switch that provides POE to it was also rebooted due to a UPS replacement

#

Well finally got the broken devices to be accessible again. Some required their own power cycling (so thank God none of them were hard wired wall outlets) and one required a full factory reset

elfin belfry
#

Ugh this is happening again. 2 more Eve plugs are now unaccessible to Home Assistant but work fine in HomeKit

#

I really want to know how to get this resolved

elfin belfry
ruby notch
#

Maybe not directly related, but it's always a good thing to set reserved/static ip's for your devices including HA in your router, and set the devices itself to automatic, hence this could resolve dhcp lease issues.

#

And always check DNS, it's all in the name 😉

#

check wireless channels not overlapping eachother, as far as possible.

#

if virtualizing from VirtualBox... there seems to be a few issues at VB update 7.2 and 7.2.2

ancient canyon
ruby notch
#

confirming the honeywell

ancient canyon
#

I usually go with Jasco or Zooz for most things. Jasco because they can control any LED fixture, where as other brands may cause LED bulbs to flicker.

#

Zooz for anything Jasco doesn't offer and sensors

elfin belfry
#

Thanks! I’ll start looking into replacing things slowly but for now I still need to figure out how to get the current stuff stable again. It seems the unavailable devices randomly switch between 2-5 of my 9 total thread devices

ancient canyon
#

I avoid bulbs on zwave, I have zigbee for bulbs and I exclusively purchase phillips hue because the offer the best colors.

#

zwave has 0 good bulb manufacturers IMO

elfin belfry
#

I used to love Hue until they started forcing a cloud account

ancient canyon
#

meh, you don't need it for zigbee.

#

just include the device and go

elfin belfry
#

In that case I just need a good zigbee box as my Hass is an ESXi VM

ruby notch
#

but unpair first in app

elfin belfry
#

Doing a full stop of the Thread network by unplugging all hubs with Hass offline. After the wired AppleTV being up for 20 minutes, now I'm powering on Hass. HomePod minis all still unplugged. If things come up as stable, I'll start slowly adding the minis back

#

I appreciate you both helping debug this. Seems like Matter add-on logs are much cleaner starting up this time. Devices detected much quicker. I'll leave it like this for an hour or two before adding back HomePod minis

elfin belfry
#

Looks like after an AP reboot I should bring up 1 border router then hass, then slowly add back border routers

dark gyro
#

Can anyone recommend a thread router node (don't need a border router -I have ZBT-1) - I n need to extend my Thread network but all I am finding are smart plugs and wall switches. The Nanoleaf bulbs seem to switch too much between router and end node, which makes my network unstable.