Currently using VMware on Ubuntu, but thinking about using a container based install instead as CPU has recently really increased for some reason (apparently Docker has changed the way stuff is done and is causing more overhead).
I see good things of people just using docker and running it raw so to speak. If i run the addons separately in docker, can i still restore my configuration from google and have it applied/picked up by those containers ?
You cannot.
Addons, as such, are a feature of HAOS (or the supervised install).
They are only a repackaging of existing applications, though, but it might be tricky to do the equivalent yourself if you are not proficient in Linux.
I guess i meant run the containers separately in docker, most addons have container equivalents.
Sure, but what I meant is that you cannot re-use addons in any way: you’ll have to build the equivalent yourself.
So, no.
Gotcha. I guess i can apply most of the configurations i used though for the addons fairly one on one? With a bit of change here and there?
The configuration values themselves will be the same, mostly, but how you pass them to the actual app will be very different, because HA addons use their own way to specify, store and apply them.
You mean running HA on Ubuntu, running as VM under VMware?
Then simplification would be to use directly HA OVA (HAOS based) in VMware. I run HA this way since >2 years now and no performance or unexplained CPU hit. Using 5%pf assigned vCPUs (4 cores on i7 CPU). And it is supervised, with all advantages.
I mean running the official OVA in Vmware player like you are. I used to have around 5-9% as well but it’s just gone up for no apparent reason.
What version vmware player are you using and what OS if i may ask ? I’m using a MacBook pro 2011 i7 with 2 cores, 16gig ram, 8 of which is dedicated to HA and one core. Latest vmware player.
Well, since I have small home lab, I have regular ESXi (free hypervisor) server, recently migrated to v8. So no underlying OS, just VMware. VM is configured with 4 cores (5% average use, 12% in peaks, that happen very rarely) and 4GB RAM (all used, but 25% active on average, 35% in peaks), so looking at statistics it seems to be overkill. 2 cores and 3 GB RAM should be sufficient, even given that I have quite large instance of HA (currently 2410 entities).
As an alternative for ESXi you could also try Proxmox. Just to find out what you prefer.