But for those who have no intention of using their 2GB Rpi 4 even for audio, never mind video, there doesn’t seem to be a way to select a reduced OS install, to make the best of it’s capabillites. Mine shows that it rarely uses more than 2% CPU, but memory is getting tight at 1.4 used of 2GB.
It only uses 3 or 4W in power generally and is well capable of handling quite a few Zigbee switches & sensors and a handful of ESP32’s providing data.
Although I worry each new update install may kill it.
While it’s great to see HA expanding to do so much more, it does leave behind those who maybe don’t want all those great new features. The whole point of HA is local control. That includes control over what we choose to use it for.
It would make a ton of sense to make the OS more modular, and allow users to choose which components to include, allowing them to choose the best hardware for their implementation.
Of course no-one with any say in the direction HA goes will ever read anything here, so we just keep buying new hardware every few years as HA’s footprint expands.
That’s a big exaggeration. If you run it on anything other than a Pi (a mini-PC for example) you can have HA running for many, many years on the same hardware. My Intel Skull Canyon from 2016 doesn’t break a sweat running HA and other workloads at the same time.
If it’s on 24-7 ‘running other workloads’, might as well use it for running HA aswell.
But a ‘Skull Canyon’ uses 10 times the power of a Rpi 4.
If a 4GB Pi was available when I bought the 2GB, I would have bought that. But ‘hens teeth’ then. I think even 4GB has a long way to go before it will struggle.
Long live the Rpi 4.
Sure, but factor in the price for renewing hardware “every few years” that was mentioned and do the math on that. Also take that you do not need to run HA exclusively into the equation.
