| 18 Nov 2021 |
| teto joined the room. | 11:02:21 |
| 19 Nov 2021 |
| happyalu joined the room. | 04:21:36 |
Maniac of Madness | Hey everyone! NixOS and Home Assistant newbie here. I'm on the unstable channel and I'm using the configuration from here:
https://nixos.wiki/wiki/Home_Assistant#Using_PostgreSQL | 14:13:36 |
Maniac of Madness | In reply to @hexa:lossy.network with an empty services.home-assistant.config = {} we are missing aiohttp-cors and home-assistant won't start Is the above still true? I'm asking cause I'm getting a "ModuleNotFoundError: No module named 'aiohttp_cors'" error in the systemd journal. | 14:15:16 |
seeMee | it works for me | 14:17:21 |
seeMee | but to get aiohttp-cors I would prob try adding auth = { }; in config part of home-assistant | 14:20:39 |
@hexa:lossy.network | add default_config = {} if you don't want to look up basic dependencies | 14:21:01 |
@hexa:lossy.network | https://www.home-assistant.io/integrations/default_config/ | 14:21:11 |
@hexa:lossy.network |
166: "default_config" = ps: with ps; [ pynacl pyturbojpeg aiodiscover aiohttp-cors async-upnp-client emoji hass-nabucasa home-assistant-frontend ifaddr pillow pyserial pyudev scapy sqlalchemy zeroconf ];
| 14:21:29 |
seeMee | In reply to @hexa:lossy.network add default_config = {} if you don't want to look up basic dependencies oh that explains why I did not have those issues as I added it because of some other dependency failure | 14:22:14 |
@hexa:lossy.network | fwiw, my personal instance runs without default_config | 14:22:45 |
@hexa:lossy.network | services.home-assistant.config = {
# opinionated default_config
config = {};
logbook = {};
history = {};
map = {};
media_source = {};
mobile_app = {};
person = {};
scene = {};
system_health = {};
ssdp = {};
sun = {};
script = {};
tag = {};
timer = {};
zeroconf = {};
[...]
| 14:23:33 |
@hexa:lossy.network | all input_* components have no deps anyway | 14:24:04 |
Maniac of Madness | Thank you both for your responses! I'll have a look at what's contained in the default_config and pick what makes sense for my setup. | 14:24:26 |
| 22 Nov 2021 |
CRTified | Not necessarily nix-related, but is it possible to have some kind of failover for HA or at least zigbee2mqtt? | 13:40:01 |
CRTified | My sd-card died, and while I'm going to switch to an SSD now I'd love to look into making the system more resilient | 13:40:50 |
@hexa:lossy.network | HAHA basically | 13:41:41 |
CRTified | So you mean running two instances in parallel? | 13:42:20 |
@hexa:lossy.network | I guess two machines connected to a postgres cluster with their individual zigbee dongles | 13:42:22 |
@hexa:lossy.network | though I'm not sure how that'd work, given they'd both like to write sensor data etc. | 13:42:47 |
CRTified | Wouldn't that result in a split brain for automation? | 13:43:02 |
CRTified | So everything would run twice | 13:43:15 |
@hexa:lossy.network | yeah, I think that's not going to work | 13:43:19 |
@hexa:lossy.network | basically set up k8s at home over 3 nodes and tada.wav | 13:43:28 |
Linux Hackerman | I think this is a case where redundancy would cause more outages than it would prevent. | 13:43:39 |
@hexa:lossy.network | and connect your zigbee dongle via usb2ip on an anycast address | 13:43:52 |
@hexa:lossy.network | sorry, I only have silly ideas today. | 13:44:01 |
@hexa:lossy.network |
- don't run your HA on a Raspberry Pi
| 13:44:34 |
CRTified | In reply to @hexa:lossy.network sorry, I only have silly ideas today. No problem, the idea on its own is probably a bad idea in the first place, given the state of sanity for home assistant 😀 | 13:44:56 |
@hexa:lossy.network | probably, yeah | 13:45:08 |