Hi,
I’m considering installing a second Home Assistant server with Home Assistant as its operating system and possibly its own coordinator – is this a viable option these days? There are also Synology NAS devices on network, which are used to back up existing Home Assistant server.
Does anyone have this setup?
For what purpose?
Thanks for any input and/or ideas!
What problem are you aiming to solve with this setup? If it’s reliability or Zigbee range there are better options.
Yes, my setup works.
Nevertheless, I asked because the topic came to mind and equipment ist lying around.
I’m always thinking about improvements.
What better options are you thinking of?
You could virtualize (Proxmox as an example), which opens up support for replicating your HA instance (to be more fault tolerant). Snapshots are also a benefit.
I would not use two Zigbee coordinators since that will give you two completely separate and potentially competing Zigbee networks. Unless it’s for a detached building or something like that I would install router firmware on the second one - that way you get a better mesh and in a failure scenario you can re-install the second with coordinator firmware and be up and running in no time.
Okay – interesting.
I once used a Novell System Fault Tolerance III server in BMW’s production line. Production would have come to a standstill if both servers had failed.
Even though I now have a “non-critical home network,” I’d still be tempted to implement something like that.
I don’t bother with replication myself even with the hardware to support it - I do 12 hour backups and snapshots before every upgrade and that covers 99% of my potential issues. But I do recommend looking up Proxmox, it’s a very powerful product. Do not limit it to HA, you can use it for containers and other virtual machines also. It’s open source so no license fees.
Another use case if you want near 100% uptime on HA is that you can migrate the workload from one proxmox-host to another (during maintenance on the host for example) - I usually don’t bother since it’s quick to do a full reboot of all my workloads, but you have the option.
