There is increased usage of ram in haos 10. Try downgrading to 9.5.
1 Like
I hope the changes you made will result in a stable instance. Donβt worry too much about the RAM usage for now.
1 Like
Thanks for your help. I did as you suggested and then rebooted the HA VM.
I guess I need to be patient now and not worry about RAM usage for the moment. Letβs see what happens in next 12 hours.
Once again, thank you !
1 Like
For what itβs worth, I think that is normal. At least on my system I experience very similar values: Proxmox showing memory as 3.5Gi used out of 4.0, but HA only showing 26% used. Iβve been running stable for a couple years, and never had any RAM-related issues.
1 Like
My proxmox shows 1 of 4GB used. It shot up to 3 or more with HAOS10 and then went back down when I reverted to 9.5
Posted about it here. Home Assistant OS 10: Better memory management and new board support - #88 by nickrout
EDIT HAOS 10.1 shows normal RAM usage of 1G. Seems to be a problem isolated to HAOS10.0.
Just a quick update : After doing Kernel and Intel Microcode upgrade and changing CPU model (host) and Disk Cache (write-through), my HA instance is running stable for 36 hours. Still under monitoring but it looks promising so far.
Iβm using HAOS 10.1 and Supervisor 2023.04.1.
@tteck and others : Thank you very much !
1 Like
Iβve setup a Ha vm in proxmox last week, but wanted setup a new one to with ttecks script.
The scripts starts but keeps hanging on the storage pool page. I can choose the local lvm but if i press ok, i stay in the same window. Same if i choose other locations. Any idea?
spacebar = select
1 Like
thanks! helped me out!
1 Like
@tteck See the post below, @seanblanchfield claims to have HASS OS running fine inside LXD. Any reason why that wouldnβt work in Proxmoxβs lxc?
I can envision issues with the nested docker using vfs storage and consuming tons of space, if it canβt be customized to use usermode fuse-overlayfs, maybe?
CMIW, In his case, LXD is being used in conjunction with other technologies to run VMs.
You cannot directly import a .qcow2 file into an LXC environment within Proxmox VE.
1 Like
Ahh I see, heβs using LXD for VMs, not lxc. Dang, thought I was on my way here.
Just a note:
To ensure compatibility with Proxmox VE on the affordable NUC-style systems equipped with N-series Celeron processors, it is necessary to install both the Proxmox 6.2 kernel and the new Intel microcode.
hi @kanga_who can you tell me how this went on the HP T520 thin client? canβt see other reports. Iβm thinking of doing the same to a similarly specβd thin client. how much ram did the t520 have?
Not sure if I have mentioned it in this thread, but I am running HA in proxmox on an hp t630. It has 16gb ram and a 256gb ssd. It also runs openmediavault with 3 usb ssdβs, pihole and motioneye with 8 cameras, Although only 2 cameras are setup for motion detection. The CPU sits at about 80% and ram at about 70%. The cpu usage is about half that it you turn off motioneye. An hp t520 will handle HA with ease.
Hi, Iβm not sure what Iβm doing wrong (or if I just need to do some other steps). I have a Dell USFF running proxmox with a single NIC. I have Unifi network gear and the connection is a trunk connection sending all VLANs. Proxmox is on the default network and my want my HA VM to sit on the IoT VLAN. The Proxmox bridge is setup as VLAN aware. Iβm using the advanced version of the script and setting the VM to be on VLAN 30 (IoT). But when the script has run HA isnβt showing an IPv4 address. What is confusing me is that my the interface in Proxmox is enp0s1 but HA is saying the interface is enp0s18. Is this the issue or is there something else I am missing?
