It wont let me enable it other devices under the parent are on. Device works at source so why is it disabling it?
Been a while since I had this issue, but I believe I removed it from the mesh and then re-added the device to the mesh.
Tried that, no dice.
What I found was the device that was always offline didn't have "hub mesh" in it's used by apps list.
However after doing as you mentioned it's actually worse and none of them have it in their used by apps list and none work.
Could be the custom parent driver that @mavrrick58 is using but I have never had a issue sharing custom devices and drivers before.
@mavrrick58 have you tried sharing your parent device between hubs before?
No I have not. I will give it a try and see what happens. I wonder what it is using to determine it is offline.
Seems to be the fact that in the device page it's not "connected" to hub mesh. But you can't add a device within a parent to hub mesh you just add the parent and everything under it gets added.
I just managed to complete a test and it was a little interesting.
The first time I tried it I got the same errors. Everything but the parent device was in a disabled status and offline. When I got it removed I couldn't get it added back. I ended up rebooting both of the hubs I was testing with and then had to add a different device I was testing with and then finally the other devices added fine. Yes they added completely and were showing up on the hub they were meshed with.
It does seem to be a bit twitchy. I am not sure why it was being so temperamental. Something else I noticed which caught me a little of guard was that the last device I added that had Debug logging still on looked like it was polling the Govee API from the secondary hub as well. That could be bad for your API Rate Limits. I don't think it is a great idea to mesh these devices based on this. It looks like it could be possible, but also looks prone to cause problems to me. Does anyone know enough about hub mesh to know what is and isn't doing when a device is linked.
That would be @gopher.ny.
The reason I was doing it was to try and of load some work from my main hub. I've had some issues with it recently since I added some stuff and matter and I'm getting hub warning over severe load. So thought move some cloud stuff over to the other hub which has nothing to do may help.
Generally speaking as long as you are not doing anything to crazy with the Govee Integration it is pretty light weight. It does hit the hub hard when the hub first starts, but generally once it gets through that initial startup it should be pretty light weight. Just make sure you have a reasonable pooling interval like 300 or 600 seconds (5 or 10 min)
The reason it is pretty heavy on startup is because it is refreshing the state values and such so stuff during runtime is available. I have continued to make changes as time goes on to try to improve it.
I think memory is a different matter though Memory is consumed as you add more and more devices. Before i upgraded to the C8 Pro my older hubs were running around 250mb right after restart. The good thing was they didn't seem to drop much though after that.
All that i totally understand trying to reduce load. I am just not sure you are actually doing that based on my observation. I really hope @gopher.ny can shed some light on if a Linked LAN API device would double up on the pooling and if there is anyrthing Hubitat suggest to control that.
Yes it's this but I have a few apps that do this a well, so it hits it hard. I need to get a pro as well basically but trying to wait to get it at a price that I don't have to pay the costly import duty.
In this case yours was the easiest to move as I only have a few devices so far. I don't think it should poll on both hubs it's just the logging gets replicated on the second hub. I'm assuming the polling is controlled and managed in the app which is on the other hub.
Either way there seems to be a issue that needs to be fixed on hub mesh.
I have had this issue too with child component devices. Tagged @gopher.ny and never got a response. Definitely an issue. I ended up doing some behind the scenes Postman hack magic to enable them for my own use.