Anyone got a NEW HA install (bare) you can test for me before I post an issue in github?
If you spin up a VM with 4-6 CPU’s, HA boots fine but if you change from 6 to 8 and then boot it stalls
Just want to see that its not my system.
Am using VirtualBox
Do you mean CPUs or Cores?
How many CPU’s/Cores do you have in your system ? , anyhow you don’t need 8, not even 6, unless you have a huge HA installation, with tons of devices, automations and scripts (and i mean Tons)
Sure you could report in Github, but if you have 8 cores, and choose 8 in VirtualBox for HA, it means it want access to all, kind of irrational approach, considering the underlying OS + VirtualBox would love to have some to
( VirtualBox + (bare) , doesn’t sounds right to me, but that depends upon what you mean by “bare” )
I do have a MASSIVE installation unfortunately. (Im running a commercial brewery from it - a 4 vessel brewhouse, 8 fermenters and 20 brite tanks).
Regardless of the size, I develop on a VM and have been transferring over to a bare metal unit (right at the moment is 6th gen 4core). I have just bought a 8 core gen i7 to reduce significantly the reboot times, hence why I upped to using 8 CPU’s in the VM (which didnt work)
Before you say its overkill, a reboot on the current machine (i7 4 cores 3.6GHz) takes about 3mins (which recently there has been an issue with the Modbus integration hanging a reboot if HA loses communication with a Modbus Hub and can take 10m to get past it) - and 10m (even 3mins) is too long if you are int he middle of a runoff from the kettle (the hops keep bittering the longer you leave them)
I choose to run bare metal on the production system rather than any form of VM/Hypervisor as it means that theres one less point of failure and one less thing to administrate.
Looks like my new machine isnt fit for purpose unless I do run a VM and reduce the core counts.
There are other means of getting a VM. eg proxmox.
I dont want to be using something in between… its just something else that can go wrong and something else that needs to be maintained/updated/backed-up
A layer of unnecessary complexity
Hence bare metal
Sorry I thought it was a VM problem.
So did HAOS(bare) work on your i7-8-cores ?, i mean not any point in trying any VM/Hypervisor , if the root-cause is that HAOS stalls , on a bare-installation with 8-core/i7
PS: im puzzled about what you are trying to show/say, with the attached Picture
Im also wondering why you believe running 8cores instead of 4, would solve this kind of communication-issues.
Regardless, you initial Post , was also full of “doubts” , as i said in my first answer.
i.e why do you ask if anyone has a bare-metal 8-cores installation, when you have issues playing around with your unspecified VM system ?
Install a “bare” on your 8cores, and report back
HAOS will run on 8 cores. At least on proxmox it will.
It most likely also will on Oracle-Virtual-Box, thou default, if you dedicate ALL available CPU’s/Cores, VB will then give the “guest” unlimited access/usage(100%), which means you most likely kill the Host and VB, so most likely he will find the “root-cause” in the logs of the Host, or VB’s logs
VB have “always” to me, looked like the most “unstable/troublesome” Virtualization you can use, thou you can configured it to not use 100%, something that a (other hypervisor) by default should/would fix
As for the 8 core machine - Ive ordered it but havent recieved it yet so I cant test it yet. I will report back as soon as I get it.
The modbus issue - yes admittedly - was a bit of a off comment. My need is really fast reboot times should an issue present itself at the most in-appropriate time (while we are working). Idle cpu usage is not necessarily an issue. From my experimentation in VirtualBox so far 1,2,3+4 cores = incrementally faster reboots. 5-8 becomes a problem and is significantly slower. Now its not to say that theres a setting in the VM that hasnt been addressed properly… which Im happy to adjust.
And I will say again… I am not interested in installing HA on ANY hypervisor/VM/docker/proxmox on the production system. It is a level of complexity not needed in a “brewery envorinment”
Also, I was not specific on my dev VM environment - my apologies for not being clear - Im runninig a Threadripper 3990x 64core 128Thread so im not running up against a VM issue.
If I can ease your concern, it seems to me that there is no barrier in general to running HA on 8 cores, and the problem must be the virtual box VM.
