2024.3.3 Are there issues known, so much trouble on Raspi 4

After updating, reolink camera’s stopped workin, SMA fell out of the sky, Buienradar stopt, Somfy integration gone, esphome slow loading 4 minutes.
Loading takes 10 minutes on a raspi 4 no twice as fast loading as said in the update announcement, for me it is a pretty terrible experience this new update.

No idea if I can do a roll back or that it is pretty much bricked.

Am I the only one who has this or are there more troubles?

Yes, just restore the backup you took before updating

Check the github of HA as this is the correct spot to track issues/bugs

Now I see 4 lovelace instances not loaded too in HACS, this new version gives me heaps of trouble. ALl was working fine previous version.

Solved after 20 minutes they seemed to have been loaded.???

New error:

Logger: scapy.runtime
Source: /usr/local/lib/python3.12/site-packages/scapy/error.py:150
First occurred: 2:42:34 PM (1 occurrences)
Last logged: 2:42:34 PM

Interface veth4622d96 does not exist!

no idea what this one means, and were it did come from it was in my logs.

Well so much for breaking new changes…

Do you use scapy? Doesn’t seem to be a official ha integration…

Maybe add the basics to your thread like yaml config:

Sorry, I think I do not understand your point. The forum is first of all to help each-other to have the best running HA instance we all can enjoy, so we do help each-other whenever we can. I know that people do not work here!!!

My question is a Home assistant related question, scapy is a part of Python part of the underlying program of all Home assistant things, so I do not get what you think you say.

Plus I do not know when you did to start using HA, but whenever you encounter a problem which you can not solve then ask in the forum, it will help others with the same problems, it will help improve things in the future and it will most likely be solved now or in a upcoming release, I use HA since the day it started, after NODO, so please Indeed.

My question has all to do with the new update of HA, since the problems started after the update.

Your message diverts all who wish to answer from the problem so please do not tell me how things work, if you have a solution please help, if not please do not pollute the topic, thank you.

All I can say, my Pi 4 runs fine, has been running fine since I bought it when the 8Gb version came out. I have no issues whatsoever on 2024.3.3

Start by disabling custom integration and custom cards, reboot, and see if the issues are solved. If so, enable one by one, restarting in between.

Did try several reboots after disabling non native HA integrations, nothing helpt. I do not have extra hardware attached to my Pi4. it is baremetal.
Can not find the reason why this is happening. I keep looking or hope to hear a viable solution or hint in the right direction. ty

Have you tried restarting the host not just HA?
I have no problem here with the latest version of HA on my Rpi4 8Gb.

Yes I did, host, reboots, total system shut downs and restarts, no errors anymore, but starting up just HA not total reboot takes now 5 minutes or more, SLOOOOOOwwwwww. Many integrations are waiting and waiting and … to load and some do not get started at all. All since last update.

What installation method of HA are you running? You say it’s “baremetal”, but is it HAOS or something else?

Also, are you running on an SSD or SD card?

Last question - can you take a look at the Host logs for any “under voltage detected” warnings?

I run HAOS on a Pi4. on a SSD for years now.

There is nothing indicating undervoltage, the Pi is hooked up to a pretty beefy power supply. I get no errors anymore in the logs, none. Still HACS complains about integrations not loading in lovelace, Buienradar a weatherforecast is gone, but worse the super long start up time.

Go To Settings > System > Repairs > 3 dots menu (top right) > Integration Startup time.

Check which integration is taking longest to load, disable, reboot, repeat until it’s stable. That’ll help narrow down what is causing your issue.

Thank you that helps a lot already, never looked in that tab, did not even know the existence of it. Tells me a lot indeed.

1 Like

That’s great indeeed! Whay does it tell you?

Another thing to note about flash storage is that they usualy slow down before they die (often turn themselfs into a read only/grid lock).

The entity for Buienradar, the weather forecast in my country is the main culprit. Taking nearly 3 minutes to try to get in contact with the service. Which is strange for it is a in HA incorporated service, not a third party any more.
Plus the fact that it no longer runs in the newer update of HA, this update killed the weather forecast service, hope someone will know why, for a lot of automations, scripts and more are dependent on this service and its entities.

Furthermore there is a system monitor that takes nearly a minute to start and a lot of other integrations taking up half a minute each.
Add those up and you surely see why things run so badly.

Problem solved, Buienradar was the culprit well it is largely, solution to it is Buienradar uses as only service IPv6, and does not get any data via that channel, so it is waiting and trying to obtain data for a long time.
Debugging option in Buienradar showed that problem.
After shutting IPv6 in my network settings in system of HA, Buienradar needed to look at the IPv4 route for data, and yes it did. Only thing I think it is new so all the names of the sensors did change after this, they became sensor.wind_speed now where they were sensor.buienradar_wind_speed.