High memory utilization after 2024.2.1 update

Hey folks, this will be my very first post here , 'til today I was able to tackle every scenario just by reading/searching around these parts…but now I admit it, I finally need some guidance

My setup

  • Core 2024.2.1
  • Supervisor 2024.01.1
  • Operating System 11.5
  • Frontend 20240207.1

Running in proxmox VM with 4GB of RAM

Here’s the issue
After upgrading to 2024.2.1 I saw an increase of memory utilization.
When I noticed that the memory was high (around 95%) I restarted my HA proxmox VM which made a tiny decrease in overall utilization, but it’s still +15% more than it used to be.

Let me show you (this is the proxmox memory utilized % for the HA virtual machine)

The strange thing is that if I check HA memory consumption, it’s low (can’t share 2 pictures as I’m a new user) , the value I see reported in Systems → Settings → Hardware → Memory is 23%

I understand how Linux handles memory, physical addresses are mapped to VMM so MMU can handle page allocations and also perform IO transactions to page cache. The main issue I have right now is: I can’t validate inside HAOS who is where because I have no common tools. (where did my memory go? who is using it?)

I’m trying to do the “go figure” but I couldn’t understand what HAOS/supervisor/integration component is consuming my VM provisioned memory.

This is where I need real help,

  • what can I do to perform a memory breakdown of the running components ?
  • commands, integrations or even config changes are welcome :slight_smile:
  • Are we just allocating more memory to page cache so we have better IO performance on this new release? Because it looks like it:
[core-ssh ~]$ free -h
              total        used        free      shared  buff/cache   available
Mem:           3.8G      871.1M      480.9M        4.9M        2.5G        2.9G
Swap:          1.3G           0        1.3G

Thanks a bunch

bump… anyone can shed some light here? :slight_smile:

I am noticing the same problem as well.

Every day I need to restart to solve the high memory consumption problem. In the “hardware” part, consumption is 20%, but in qemu it shows consumption of 95%.

If you are not using 2024.2.1, then this is not the right topic for this question. But then again, due to the way Python memory management works, you cannot see how many memory is actually in use from the outside. So if you actually experience problems, create a new topic to describe the symptoms, but If you are just worried because you think memory is 95 and reboot because of that, stop worrying.

1 Like