Hassbian - DNS lookups fail for ~15 minutes, chronically

Just spent a few days messing with Hassbian, ported some of my lights/sensors to the architecture and works a treat - impressed

However, every couple of hours, for 15-30mins, i get getaddrinfo() -3 failures when polling sensors and lights (from dev info/about/log). Service calls (over http) may fail during this time, so scripts/automation are unreliable. It’s running over internal wifi (rpi 3B)

HA 0.76.2 - all my py use http.client, would porting that to requests/asynchttp improve the situation?

I’ve tried paring it back, and adding (my) components piecemeal, but it doesn’t appear connected to my components - any ideas please?

getaddrinfo, at least to my knowledge, should have nothing to do with your components, scripts etc… This error appears when the machine has connectivity issues. Is it possible to, at least temporarily, connect the Pi via cable to compare if bad WiFi reception might be the issue?

thanks Daniel - i’ll get a hub and see if that makes a difference …

i’ve made the rpi wired instead of wireless, and the dropouts have ceased (24 hours without a single one)

Is wireless known to be flakey? i’d rather resolve that than start bridging things so hass can find my devices (on the same lan)

The only impact I have seen is if you are trying to use Wifi AND Bluetooth. BLE is even more problematic in my opinion.

ah! that’s exactly what i’m doing - BT and WiFi

thanks for your insight

bridging …

My recollection is that the Pi uses the same device for both the Wifi and BT which might contribute to this challenge. Besides… Ethernet just works… no ssid and credentials to deal with. I use BT (not BLE) for presence detection on my pi and it is flawless, but I use Ethernet as opposed to wifi.