in my effort to streamline my energy monitor system to an all in one device to reduce clutter. I think I finally got it. it now does everything… I used the bpi-r1 and bpi-r2 s but they are never quite enough to do everything.
I got it done with intel router board that has 8 gig of memory 128 gig msata and a 500 gig sata drive ( it did not have sound so I added a 5.1 usb audio to it )
it handles everything my other energy monitor could ( zigbee@mqtt hub, home automation, data collection and base router system) but now it also acts as computer desktop and entertainment system (kodi) that can be voice controlled via the mycroft AI or by remote control… now just to get my head around how to build gui skills for mycroft AI and I will add more functions to it …
if you want to attempt to build a full house single unit from one of these router boards - the hardest thing I found was setting up this unit was to get all the network ports to play nice. but in the end the simplest way was just to use the network manager gui ( nmcli) and then use “share to other computers” then your lan ports will use 10.42.0.0 IPs but you can change them by using this command
nmcli connection modify local ipv4.addresses 192.168.42.1/24
( local being the name of the connection)
the nicest thing is you can use this command to give each shared network port their own network range
so from my little effort i pulled one router ( router and energy monitor), one ai unit and a desktop computer and combined them all in to one with an energy foot print of 10w ( the same as each individual devices energy foot print)
Interesting solution. My issue with this approach (putting all your eggs in one basket) is that if it goes, the whole lot goes. I have suffered before where upgrading one part, upsets other parts. I’ve said before, I have an old Laptop with ProxmoxVE installed. Each component runs in it’s own container which I can snapshot before changes and backup every day. The bonus; it has a built-in UPS.
not really an issue. ( I use VM all the time and they are handy ) but in this case where resources are limited I do this --once you have you have the base working the way you like it simply image it . and after any significant upgrade simply image again. best method I find build your OS first then once done link your home directory and a data storage to a separate drive each on different partitions . that way if the OS dies due to hardrive failure or other catastrophic event simply re image to new os drive from the last backup. plug in the home drive again and you are up and running again in 30 - 40 minutes , if the home drive dies well then you are in the same boat no matter what .as a secondary and if you are handy simply partition the home drive and have the main OS back up to it except for fstab ( as that needs a different ftsab) , and in event of catastrophic of the OS harddrive you can simply point it to boot at the secondary drive and you are up in 30 seconds . easiest way to do this is when you setup the home drive install linux on it first but with a / , /home and /DATA partition and simply have your backup software update and add files to this OS root partition as your running main OS partition changes ( basically operating as a software mirror raid) … personally for years I never being down for more then a 2 hrs after a catastrophic event - basically what ever time it takes to load a new copy of os on a hard drive and the software I use . then simply plug in the home directory drive and I am back to exactly the the way it was before the failure, then every few years image the home directory to a new hard drive and replace the old drive I got into that habit when using opensuse as that is how it always installs Linux as a / and //home partitions … ubuntu sucks in that respects as you have always have manual install it that way
