thank you, i’ll give it a try. my main concern is to get my connection to fhem working again.
updating the connection of my wallbox will be not a big deal, but i’m afraid of changing anything related to fhem (as i rely on it to control my somfy rts blinds with that (with a nanocul, which is not supported by homeassistant directly, sadly)
how should/can i ensure that?
Start with running MQTT outside HA. Follow my guide which hopefully answers all your questions. Then let me know how you get on and where you are stuck, hopefully nowhere.
Make certain this setting is applied to your HAOS VM. It is under the storage tab, bottom right.
OT: this forum is annoying. eveytime i get this shitty error. just want to post normal text.
select don’t show and it should never ask you again; useful for some of us who want to post code
Yes it is the default but some people think they know better and change it. Always use the default for Async IO.
so, i accepted your challenge
mqtt is now a seperate lxc and i managed to connect fhem & homeassistant to exchange device-states.
took me a while, although i had to make sure to not destroy any working setup
Happy it worked for you as it should for everyone else. MQTT is relatively easy.
once, you understood it, it’s fairly easy - true.
will see if all my work pays off now and my HAOS-VM works without too high memory in the future.
thank you again and merry christmas!
take care and Merry Christmas and Happy New Year free of HAOS problems.
Again with the internal fullbackup of HA, memory increased to the limit and does not go down anymore. Before the backup memory usage slightly but steadily increased. Totally same behaviour as on previous VM.
Sadly the changes didn’t help:disappointed_relieved: (besides practicing proxmox )
First drop was a reboot of VM, the other small 2 a restart in HA. No it’s at the memory limit.
Couldnt remember this behavior in older versions. Just since a while now (not sure if it started just with 2023.12 or already 11)
There is something in your previous VM, which breaks things. I suggest a new VM and starting from zero. Now your MQQT is outside hopefully will be less work to do. Something breaks your instance and you are the only one who can find out by testing things one step at a time.
You could also share what things are installed on top of a vanilla HA install. e.g. HACS integrations, add-on, HA integrations, anything else which affects configuration (e.g. anything custom in configuration.yaml).
You mean to start with a new VM and new disk and import a backup or completely start from scratch?
2nd would mean a lot of work and I’ll not do that, then I’d keep the high mem and let it run as long as its working.
I think my setup does not contain any very special integration or so…
I suggest you remove the Google Theme and see if issue gone. The rest seem OK.
My suggestion is to create a new VM for HA. Then leave as is to see if HA over a couple of days has mem issues. Then add 4-5 integrations into HA then test for a couple of days. Then add a couple of HA add-ons, test for a couple of days. Then add a couple HACS integrations, test for a couple of days. And so on until you find something that breaks HA
got it.
i’ve removed google theme for now.
when there is time i’ll start another, new HA-VM and see what happens.
thank you
just running “core check” on cli lets the memory peak to the limit and it does not go back after finishing the checks.
the drop was the reboot of whole vm, then it was at 1.35GB, started the check and it’s back up to the limit since then.
i’ll leave it for now and go for a run
hey,
i’ve created a completely new VM for HA and started from scratch. before installing anything or migrating i just did the update to HA Core 2023.12.4, this caused memory increasing and even after finishing the update the memory usage stays at the same limit as my other HA-VM, so not sure if anything inside my HA causes an issue, as the system is “naked” now. (nothing installed, added, created just 1 user to log in)
my original HA-VM - interestingly totally same memory usage
both machines are running exactly the setup that you suggested.
does this sound right for you that memory increases and stays high after finishing all tasks?