New to HAOS.
I installed Home Assistant OS 18.0 now two times on proxmox (latest) as a VM and everything works. But if I change the RAM-Size, for instance from 2 GB to 4 GB, the VM will not boot and gets a yellow exclamation mark! Changing it back to 2 GB, it is working again.
First time was with the helper-script from community-scripts.org. Second time I did everything by myself.
Another user reported to me, having no problem with his (older) installation of HAOS and changing RAM-Size in proxmox.
Not sure where to report my problem, so I started here. Also HAOS is the only VM I am running, everything else is LXC...
That is not the point and was already mentioned in the start-post. If you want to contribute, make a new VM and install fresh on the latest proxmox with the latest kernel.
Jun 27 11:53:01 pve pvedaemon[14772]: VM 451 started with PID 14783.
Jun 27 11:53:01 pve pvedaemon[1744]: <root@pam> end task UPID:pve:000039B4:00005DA4:6A3F9D7C:qmstart:451:root@pam: OK
Jun 27 11:53:07 pve QEMU[14783]: KVM: entry failed, hardware error 0xffffffffffffffff
Jun 27 11:53:07 pve QEMU[14783]: EAX=00000000 EBX=bc003d28 ECX=000000b2 EDX=000000b2
Jun 27 11:53:07 pve QEMU[14783]: ESI=00000000 EDI=00000058 EBP=bc003c90 ESP=bc003c68
Jun 27 11:53:07 pve QEMU[14783]: EIP=00008000 EFL=00000002 [-------] CPL=0 II=0 A20=1 SMM=1 HLT=0
Jun 27 11:53:07 pve QEMU[14783]: ES =0000 00000000 ffffffff 00809300
Jun 27 11:53:07 pve QEMU[14783]: CS =f900 7bff9000 ffffffff 00809300
Jun 27 11:53:07 pve kernel: kvm_amd: set kvm_amd.dump_invalid_vmcb=1 to dump internal KVM state.
Jun 27 11:53:07 pve QEMU[14783]: SS =0000 00000000 ffffffff 00809300
Jun 27 11:53:07 pve QEMU[14783]: DS =0000 00000000 ffffffff 00809300
Jun 27 11:53:07 pve QEMU[14783]: FS =0000 00000000 ffffffff 00809300
Jun 27 11:53:07 pve QEMU[14783]: GS =0000 00000000 ffffffff 00809300
Jun 27 11:53:07 pve QEMU[14783]: LDT=0000 00000000 00000000 00000000
Jun 27 11:53:07 pve QEMU[14783]: TR =0040 691ca000 00004087 00008b00
Jun 27 11:53:07 pve QEMU[14783]: GDT= 691c8000 0000007f
Jun 27 11:53:07 pve QEMU[14783]: IDT= 00000000 00000000
Jun 27 11:53:07 pve QEMU[14783]: CR0=00050032 CR2=bea01000 CR3=0012d000 CR4=00000000
Jun 27 11:53:07 pve QEMU[14783]: DR0=0000000000000000 DR1=0000000000000000 DR2=0000000000000000 DR3=0000000000000000
Jun 27 11:53:07 pve QEMU[14783]: DR6=00000000ffff0ff0 DR7=0000000000000400
Jun 27 11:53:07 pve QEMU[14783]: EFER=0000000000000000
Jun 27 11:53:07 pve QEMU[14783]: Code=00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 <bb> 4d 80 2e a1 38 fb 48 2e 89 07 2e 66 a1 30 fb 2e 66 89 47 02 2e 66 0f 01 17 b8 08 00 2e
I then enabled set kvm_amd.dump_invalid_vmcb=1 to dump internal KVM state and fed that output in an AI but it became to creative to my liking what to do next so I stopped. Will run the VM with 2 GB, which is enough and hope for a fix in the future. But good to know that it is not related to HAOS.
I checked the RAM with Windows Memory Diagnostic, it is all good. PVE has 12 Gigs RAM but runs on top of Hyper-V, which has a lot more Memory. But again, I am using only LXC and PVE has enough free memory. I even gave it more, doesn't change anything. After all, it is a me-problem.
I'd create another VM with HAOS and more memory. Then create backup on old one and import it into new one. Finally you change IP on old HAOS, stop it and set old IP to a new HA VM, so that old IP will work on your new system. You loose about 15 mins of history, that's all.
2GB od memory will be quickly too low when you increase your system. On my system on average HA uses just above 2GB, and it can go up to 3 from time to time.
Mine is using 0.7 GB according to haos, so I will not do anything. It would be interesting though, how a new vm with more RAM from the get-go would behave here...
So you are running HAOS on a three-layer stack with two stacked hypervisors??? If so, you are likely to be the only person on Earth doing this, and maybe not coincidentally, the only person on Earth to have this problem.
A more reasonable setup is to run Proxmox on bare hardware.
What you have is "windows -> Hyper-V -> proxmox -> HAOS" and you might be the first to try this. My bet is a bug someplace in the fist two layers, either Windoes or Hyper-V because the Error that Proxmox is shows says "broken hardware".
If this instance of HAOS is runing a real house and not just a set environment you really need to reduce the dependancy stack or you will be taking down things like lightbulb and haeting and door locks in the house just because of some rndom surcuity update in Windows. You expect lights and such to work for like 40 years but the taller the stack HAOS runs on the less realiable it because , as you have found already.
I'm having good luck using a repurposed Mac from 2014. It has an Intel i5, 8GB RAM and small SSD for under $100. I replaced MacOS with Proxmox. The system is actualy "overkill"
Workstation and Server Pro versions of Hyper-V are based on the same Type 1 Hypervisor. Server versions offer features not included in the Workstation versions. Windows Home doesn't include Hyper-V at all.
