Today I want to tell you the moment of panic I’ve had these days with Home Assistant. Turns out I was so quiet in Valencia for the last long weekend and I saw that the devices that get me home from Home Assistant were unanswered. I went into Home Assistant and it seemed to be all right, but I had an alert from the last backup, which you know I used the add-on for Home Assistant from Google Drive Backup which is a real wonder… but the fact is that in this add- on there was a mistake because the last
copy had given a mistake… and it was not a Google connection error or that I would have run out of Drive space… it was an I / O error, an Entry / Exit error… and that’s very scary., usually means there’s a mistake in the disk or the controller or something… in any case, Since I’m an idiot, I thought a complete restart could solve the problem… and none of that. The Nuc stopped responding and never restarted.
The problem of high availability
As you know, I’ve always said that What I don’t like about Home Assistant is your lack of a high-availability configuration.. In the end everything depends on the system where you have Home Assistant installed that can be an old PC, a NAS of those that some have at home because you’re a pirate… I don’t know how to swim that has a NAS that uses it to backup only from its photos and videos, but well… the case is that Home Assistant can also be installed on a Raspberry PI or a NUC… but yes, only in one. You cannot have a Home Assistant installation on a Nuc and a PI or on the NAS and another device for which
e, in case of failure, reboot or whatever, your entire system is still working. There are some approaches there of people who put the Home Assistant database on a SQL database and this database stores it in some NFS system to be accessible by two facilities… I think it’s great, but in the end the Zigbee or SkyConnect will have it punctured only one of the two. What if one fails, you just have to move the zigbee to the other and the devices will detect that the physical direction of the zigbee that is like a unique MAC address is the same, so they will connect without knowing what host is behind, where is it?
♪ I’ve got that prick… ♪ but that includes manual work.
In HomeKit however it is not so, if you have devices connected directly to HomeKit by wifi or by Thread, even if a HomePod is disconnected, if you have any other, it’s okay. If you are damaged by the Apple TV but you also have some HomePod at home, the system will ‘balance’ the device centre from one to another. transparent for the user. It’s clear that if you get the router you are fucked, but even in that case you can have several mesh routers as is my case with the Eero and would also take control of all wifi devices if the other one is damaged.
Obviously if you have a bridge, it’s not like that, if you have a Hue bridge and it’s broken… because the light bulbs run out of connection. Besides, if you buy another one and you have 2, you can’t match all the light bulbs with both of them so that, if one gets damaged, the other takes control… each one has associated light bulbs and they don’t share them with each other.
This is basically what happened to me.. When the NUC stopped responding, all the devices that come to HomeKit that appear to me in the Home Home App but that come there from Home Assistant, they stopped responding. 35 or more or less. The Matter devices did not know and could be managed directly from the Home App Like the native HomeKit, but all the zigbee went to hell.
He wasn’t dead, he was a parranda.
The fact is that when I came home after a few days of silent panic from Valencia (children, learned lesson, do not restart Home Assistant if you are not close to giving it a shot or watch that all starts as it touches), I connected the NUC to a monitor and I saw with disgust that it did not start, a message that no bootable unit had been found.
–
The first thing that I thought the record had cascaded., even though it’s a good SSD disk. I’ve looked at it being well connected and then I’ve started the system with a USB bootable with Linux that I have for… well… for my things, and I’ve seen that from there the album was perfectly seen, I recognized all the partitions, the checks said it was all right… but it didn’t start, it seems that something had happened with the boot partition and the system recognized the partitions but did not detect any as boot
ong >. So I I downloaded the image of the Home Assistant OS operating system and installed it from scratch. Then I started with Home Assistant completely newly installed virgin, downloaded the Google Drive Backup address-on and restored backup… and by reboot I already had it all running back smoothly.
The process itself has cost me less than 1 hour. Once the problem has been identified, it’s all been pretty fast, but the uncertainty of not knowing why it happened is something that really bothers me. Who tells me it’s not gonna happen again tomorrow… or next week… and you can’t do anything. Yeah, I could have another Nuc ready for you to just move the zigbee device and restore backup as I told you before, but it’s not an automatically much less solution.
So This has served me to reaffirm the idea of continuing to embrace HomeKit., to renew the devices little by little by little by Matter devices and to flash the Shelly with HAA in order to be able to remove Home Assistant devices or, at least, to have HomeKit publish them on Home Assistant for having it on 2 different platforms and to be able to test and not the other way around. It’s true that it only took me one hour to repair… but the cold sweats and fear I’ve had in the last few hours every time my wife approached a switch, I don’t want them or my worst enemy. wp: block {«ref»: 1612} /