… and have as happy Easter as possible.
Good catch. True for me as well.
I don’t see how having integrations added via UI in a .storage directory changes any of that. I add an allergy sensor, I provide my location, it’s done. There’s nothing to check because it is already valid. It is stored as json and I can back it up as usual. I don’t need to share the code for adding the allergy sensor with the community in order to show how to set it up or make it different because we all set it up the same and end up with the same results. There’s nothing else to change because all the options are already provided. If you want to add more options or change the flow then it changes for everyone from the integration itself and that is where the truth is.
What I do with that sensor afterwards, how I use it within automations or scenes or scripts…anything that would actually be shareable is still there for us to yaml away in. You can promote those changes, it is still config-as-code, you can find and use yaml from others or be inspired by one to apply it elsewhere.
Has anybody else seen any memory issues since 0.108.0?
It goes to the point where my RPi4 becomes basically non-responsive and trying to e.g. access the Configuration panel results into an error message.
While 0.107.7 runs on 125MB of memory the homeassistan container went up to 2GB over night:
Tried with all versions since 0.108.0 and am seeing the same issues with every single one of them.
What’s your setup ?
On a pi3 memory usage is at 290Mb
Same config on a slightly smaller footprint pi4-4 with ssd is 506Mb so it may just be using what’s available (why else did you buy extra memory )
Not sure though, any other people notice anything ?
Using about 302MB on my Pi3B… looks normal to me.
It’s set up as a ‘Supervised Install’ on a Pi4 with 4GB - so far it’s worked great.
And it does again as soon as I go back to 0.107.7.
Same with the load values, they have gone from way beyond 1 back down to below 0.5.
Wow that’s pretty bad.
On 0.108.3 pi3b I’m on 0.13 (5m)
And on the pi4-4 on 0.03 (again 5m)
16:46:36 up 39 days, 3:13, 2 users, load average: 0.64, 0.58, 0.63
on my Pi3B, I have a lot of integrations going.
I have problems with Broadlink as well. RF Lights work fine but the rollers (defined as switches) won’t work. Something interesting happened though, as one roller gets only the open instruction but doesn’t close. The code for opening the roller is at least a third in length. Is there a new limit to the code length?
Read the whole thread, and the linked issues.
TL;DR it is broken, it is being worked on, and there is a workaround.
Interesting… I made experiment copying unchanged mini-graph-card configuration that works and does not work between cards and found strange behavior. Card that work perfectly as stand alone card, copied to card where graphs are embedded into stack-in-card continues to work as intended. Card that does not work within the same stack, when copied to outside as stand alone card started to work as intended. You can see this on screenshot below:
- Upper left graph is exactly the same one as stand alone one at the bottom (so working fine as such, but not when embedded)
- right-most upper graph is copy of stand alone chart that work perfectly elsewhere, and continues to do so ahen embedded into stack
This leads me to conclusion, that perhaps there are specific mini-graph-chart settings, that leds to such behavior. I’ll try to play with these to see if some conclusion can be drawn.
You should raise an issue with the custom card developer or discuss it in the mini graph card topic.
Hi I’m also having the same issue. anyluck on your troubleshooting?
Have to tried clearing your browser cache?
god damn… i’m running on VM server… normally takes seconds to restart, now it’s more like minutes.
i stupidly updated both supervisor and home assistant at the same time, now I don’t know what’s the issue.
Since the update to 108, currently on 108.3, I have noticed that I get continously logs with:
[homeassistant.components.device_tracker] Updating device list from legacy took longer than the scheduled scan interval 0:00:12
Anyone have an idea on what has happened? Never had this before
Running Hassio on Rasp 3b+
Since 107.7 the update notification in the GUI does not work correctly. The updater services reports the correct new version, but I did not get the update button showing up. I did the update via cli and everything works fine. But now it’s reporting the wrong latest version in the gui and the cli as well.
How can I fix this?
config $ ha core info
arch: amd64
audio_input: null
audio_output: null
boot: true
custom: false
image: homeassistant/qemux86-64-homeassistant
ip_address: 172.30.32.1
last_version: 0.107.7
machine: qemux86-64
port: 8123
ssl: true
version: 0.108.3
wait_boot: 600
watchdog: true
Have you tried clearing your browser cache?