After updating to Supervisor 2026.02.2 this morning, the system is sluggish to the point it disconnects from the network. A few python3 errors on the console related to running out of memory. I’ll try to roll back but there are new integrations that I cannot proceed on until the Supervisor is updated. What is going on and do I now need to buy new hardware to run the server? Currently running natively [haos] on a laptop pc for over two years, not a docker instance.
Installation method Home Assistant OS
Core 2026.2.1
Supervisor 2026.02.2
Operating System 16.3
Frontend 20260128.6
EDIT: More info on the hardware:
AMD A6-5200
4Gb RAM
700 Gb HDD SATA
Here is a couple of the out of memory errors: (these are all hand typed as I cannot reliably connect to the server and a restore attempt will NOT overwrite the new supervisor).
[numbers] Out of memory: Killed process 7365 (python3) total-vm: 3006392kB, anon-rss:1469936kB. shmem-rss:0kB UID:0 pgtables:8968kB omm_score_adj:200
[numbers] Out of memory: Killed process 7841 (mass) total-vm:5223900Kb, anon-rss:2137528kB,file-rss:5732kB, shmem-rss:0kB, UID:0, pgtables:8968kB oom_score_adj:200
This goes on including core dumps and the like…
I tried to rollback to the backup prior to the Supervisor upgrade but it will not overwrite the new Supervisor. Do I have to start over from scratch?
Resolved by rolling back a full month (more recent backups did not solve the issue). One should note that the Supervisor never gets rolled back, once it gets updated, it stays that way.
Memory is a bit higher than it was before the updated Supervisor, but back to a place where the system will actually run. Unsure where or what caused all the commotion but at least I have a system again.
Yes same here - starts with 6GB of free memory and in course of 1 or 2 minutes its gone.
Havent found the source of it yet.
It also happens in safe mode. Maybe its linked to the overview page - it does not show even when there is still memory left. Downgraded to 2025.1.3 and the problem remains, which I did not expect.
Then I renamed my db file to start fresh, to make sure it was not a corruption of this file of some kind.
Then I disabled my recorder mariadb connection in configuration.yaml. after that memory usage seemed stable and I saw a new db file being created.
After that I updated to latest (2026.2.3) again and no memory issue present.
I am not sure what cause the database to get in trouble. If it was a migration during upgrade or if it was not caused by the upgrade at all.
i may need to be a bit more restrictive to limit recorder in my larger installation with many entities - and need to see what database settings can prevent these troubles in the future.
So - you can try to rename your db file and comment out other recorder configurations to see if that resolves it.