So it looks like I might be on the edge of C8 capabilities I'm running out of memory. 
I have read the migration doc but there are hundreds of threads with varying solutions and issues.
Just looking for a basic load distribution suggestion.
I have Matter on now with one device. 100+ pistons and probably 50 Z-Wave and 67 Zigbee.
Is that considered loaded?
If so is it best to just have the C8 do radio and the Pro do the WebCore and all the apps?
Not sure what segment creates most of the load.
What makes you say you're running out? What level is the free memory at and what is the uptime of the hub?
I'd say not. There are plenty of users on here running way more devices and apps than that on a single hub.
Free memory drops over time but it usually levels out. Mine takes a dive initially (from about 500,000 at start up) and finally hovers around the 200,000KB mark. It's around that level now after a fortnight.
My memory is usually a stable 200 Mb.
I am a beta tester and can't test some of the new features without low mem warnings.
I can't say anything more about the beta.
I find that even at 200, during a regular update, I will bottom out at 50.
This is the level where my auto-reboot piston kicks in suspecting in imminent failure.
1 Like
I just saw your other post in the beta forum - I'm holding out for the C-10 Turbo!
I think you should focus on what is creating the load. I have no load or memory issues on my C8.
I have 31 Zwave Devices and 72 Zigbee devices. In all I have 347 devices.
I used to run about 200 pistons in Webcore, now that is all written into custom apps. I was running the 200 pistons on my C7 without issue before migrating to the C8, and I have added a lot of stuff since the migration.
I keep the hub very busy all the time, and things still run very quickly with things I see, such as my lights coming on fast with presence from my scene apps, even with tons of stuff constantly going on in the background.
I keep device and app logging to minimal unless I need it, but I do push a lot of stuff out to Google sheets constantly for logging.
I think I saw a low memory warning once, and it went away.
I just don't think you need a second hub, I'm pushing my C8 pretty hard and it is going strong. Maybe one of the custom apps you are running is doing it? You may have to turn on some debug logging in your apps and drivers one at a time, and watch the logs.
My hub runs just fine with an avg load of 20%.
It's the upcoming features I'm testing that kill my memory.
Really can't elaborate as I'll get me beta hand slapped. 
Ah. I read it as it being your normal hub load based on the original post.
Yeah, confusing.
I can't really spell out everything yet to make sense.
My recommendation would be:
All Z* devices on one hub
All automations & Matter on the other hub
That is essentially my setup (except I have more than 2 hubs…) and it works great. My Z* devices hub is a C8 and has not indicated any resource issues. (135 devices total)
So a plain C8 stripped of WebCore/Apps and Matter is memory friendly for the new builds.
Good to know.
Thanks
1 Like
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.
