I didn't know that, but in my case as I'm passing the device over Maker API to Node Red where all my logic is, if I delete a device in Hubitat and then re-add it, I have to then put it back into Maker API and then update any use of the device in Node Red. Unless I've been missing a trick all along?
Just a further update to this - I have now moved all of my Zigbee devices across from Hubitat to Z2M.
I did initially have all of them on a single Z2M controller (53 mains repeater devices and 40 battery end devices) and while it worked I did seem to run into some issues, in the same way as I did on Hubitat, with that many devices on one controller. Also with that many devices it took forever to load the network map and it was very hard to read it when it did as it looked like an explosion in a pick up sticks factory. This backs up my theory that Zigbee networks generally seem to run much better with a max of only around 50 or 60 total devices on them.
So I setup another Z2M network and controller on another Raspi. Using a different Zigbee channel and changing the top level MQTT code to zigbee2mqtt2 it feeds seemlessly into my Node Red control system. I've ended up with 53 devices (32 mains 21 battery) on the master hub and 40(21 mains 19 battery) on the secondary.
It has been running like that for several weeks now without any issues at all, and everything is running superfast. The network maps for each controller load quickly and are much easier to read than when everything was on one.
I am now totally happy with this as my home automation solution. My rules have been on Node Red for a long while so now to have all my Zigbee stuff on the same infrastructure system makes things much cleaner.
If a raspi should fail (not that they seem to) I can just pull another one out of my spares box, plug in the existing SSD, change the IP on my router and everything is up and running again. I've also got automated backups of the data directory for each zigbee2mqtt instance so I can replace a Zigbee controller if that fails, or I want to upgrade it in the future, and get that back up and running quickly too.
I have had a good run with my Hubitat's, having started with one of the early C4s, and this community is brilliant, but I've just (whisper it and sorry!) sold both of my C7's on Ebay. I'm now convinced the Raspi/Node Red/Zigbee2MQTT solution is better for my needs.
I would suggest you install on some kind of SSD attached via USB. Sd card fail fairly quickly compared to a small SSD. The newer firmwares support bootijg from USB natively.
2 Likes
I think he mentioned that, but totally agree, anyone running a pi 24/7 should always be using an SSD.
As an alternative you could also explore using HA running virtual on a PC/NUC/thin client. You can then have Z2M, NR and HA all running together. If there is a failure simple restore the virtual machine! Also I've found fanless PC's to run way cooler than my pi4, and have a lot more headroom to add things like camera recognition etc.
2 Likes
Yes thats what I'm doing. Both Raspis boot from SSD via USB 3.
Maybe that's for the future, and HA has been mentioned before, but I can currently do anything i want without it. Both raspis are using less than 1% average load so i think i still have plenty of headroom. The main one is connected to a relay board that directly controls my heating and ventilation and that one is currently showing a processor temperature of 39.1 Deg C. Oddly this is quite a lot lower than the secondary one which is currently at 49.6 Deg C. So I guess the relay board must be acting as a big heatsink via the header pins. I have no other heatsinks or fans fitted to either, so if they did start running hotter I could add either or both to them.
1 Like
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.
