thats not that serious bud imho, the memory issues others were getting litrally bricks it, I could not even grab the logs through the console (monitor and keyboard hooked up to server). it would hang that fast.
sure, Ive had that with an endless automatic loop onceā¦
yet, the slowly but continuous creep has been bugging us for some time now, and its not ok.
just realizing Mill is at it again:
Error requesting mill data: [Errno 104] Connection reset by peer
and that integration often throws errors and updates those everyn15 seconds or soā¦
. Maybe its causing this creep too.
You guys will have to narrow down the problem and report it. I have a NUC, running fine without issues. No memory creep, no jump to 100%.
The memory creep is most likely an integration. As for the jump, you can add samba share and watch the logs from another PC to see whatās happening.
well, there was this and a separate it was closed as not enough info. but we dont have more infoā¦ so its a catch 22 there.
Maybe another integration is causing this too, and as said, Ive seen the jump to over 20% default since the Bluetooth integrations got active.
other than that, I have no serious lead besides things like:
Logger: homeassistant.components.mill
Source: helpers/update_coordinator.py:151
Integration: Mill (documentation, issues)
First occurred: 11:22:35 (400 occurrences)
Last logged: 13:37:12
Error requesting mill data: [Errno 104] Connection reset by peer
Timeout fetching mill data
which wont help. although, the higher memory is also present when Mill does behave. So not sure that is interfering really, but an issue is raised
thing is, this is not release specific. But memory usage does get higher with each release it seems. slowly, but steadily.
Just like the number of entities 
And thereās not much any of the devs can do without a solid lead. So repeating yourself here with no information isnāt really going to spark change.
You want them to look for a needle in the haystack in a field of haystacks. Itās just not going to lead to anything.
Since you mentioned before that you have a test instance, could you remove integrations until the issue disappears (assuming itās an integration)?
yeah, aware of that.
but it could be the other way around. As I now have only 1 instance I can only report that. many of us could produce more. thatās why I mention
no longer, but still I did remove, and reported results there yes. thereās only so much removing integrations can do, or stripping down unto a bare instance will result in. ofc the problem is of reproducing under specific circumstances and pinpointing those. stil, been there done that. spent a lot of time for the project.
and yet, the numbers keep growing, its a generic trend.
That Zwave memoryleak problem mentioned above is sorted out and no more actual. At least, most of the affected people inclusive me rapported it done.
1 Like
Keep mentioning the issue every now and then. Hopefully someone else will have the same issue and they can add what they have to figure out what is causing it. I had an OS issue that lasted over a year before anyone was able to figure out what caused it and a solution. Many times āidentical hardwareā ( not identical due to manufacturing variations) will work differently with the same software. My issue was a timing issue that effected a small group of RPI 4 users.
Just restore backup before i had the issueās. I went back to 11.2.
Update form 11.2 tot 11.4. I also update a few other. Status is this after restore:
I will leave this state for a day and then start updating 1 item to see where it go wrong
although I havenāt really noticed issues with the memory, I had a look into my systems history and noticed the following:
Unfortunately, the history does not go further back - due to my current default recorder-settings, but you can see, that the memory was more or less steady during until 16th / 17th of November.
After that, there are peaks and drops but no obvious situation that would indicate a memory leak or something like that.
The interesting part here is for sure, that 2022.11.03 was installed on November 17th - 4 minutes before 6 PM - which is the first occurance of the drop in the history graph.
Maybe, that could help a bit?
Hi hoomies one request please dont be mad this is in good faith please stop with those updates they are too frequent i dont like them it took me some very precious time to set everything up and after every update they are ton of fixes i dont have so much time to keep up and cant i barelly got this to write.
Before you continue please make something usefull like some kind a cleaner button so it truncate those loging files. My instance is constantly running out of space and have to reboot it. And have to keep amonitor connected to clean it i hate this. These updates made me quit nabucasa it didnt make sense anymore. So please slow down sun will come out tommorow too.
And i have a supervised installation and no im not having any backups localy and hdd is sss 256gb which gets cloged and yes there are no addons just esphome ssh samba thats it
You donāt have to install the updates. 
I like them. Beside the little (but very frustrating) changes in this year, which I completely not like.
If your logs are filling a 256GB ssd then you are doing something wrong. Start a new thread re that, after you have looked at the logging docs.
If you need to use a monitor to truncate a log, again you are doing something wrong. Use ssh.
4 Likes
If itās really his logs filling up 256GB, then something might be spamming them with a repeated error at a very high frequency. Should be very easy to identify the culprit by simply looking at them.
Oh and happy cakeday lol
Exactly, and thanks. 6 year user, 6 year addict.
3 Likes
Yes i am time is my problem cant fix all at once so logs are building up thats it.
Thats not quite helpfull you cant choose what to update and what not you update whats given and again and again i update mostly for security reasons and sometimes you get it and sometimes you dont.
If there is a security issue with HA, then you will see a big banner about it on the forum as it was last time.
You donāt have to update, you can run as long as you want an old version.
Currently you can also disable Supervisor updates if you want.
And the others are right. If you have 256GB SSD and that gets full and you need to plug in a monitor just to clean up log files for space, then you are doing something really wrong.
It might be not helpful, but honest. Open another topic and ask for help if you have this kind of problems.
3 Likes
New HA minimum hardware requirement: 4Tb drive, to allow for insanely spammed log file. 
Sorry, I had to.
5 Likes