I’ve been running a supervised install for years now and finally migrated to HAOS, however, it’s been extremely frustrating that this is mostly a read-only FS and theres no package manager.
So a few things, I see that there’s a few mounted dirs that are writeable
What would you guys recommend for putting some statically compiled binaries like lazydocker, for example?
Second, are there any tips about getting some extra packages on there? I’m really missing vim, htop, lazydocker like I mentioned, and a few others. I know vi is there, same with top, but those are shit I see some folks mentioned apk (the alpine package manager) in other posts, but that’s not in the host HAOS instance, right, that would be in one of the containers which is not what I’m looking for.
Anyway, hoping to get some more control over my HAOS install - any tips would be appreciated!
HAOS is locked down and not meant to be tampered with. Supervised installation was also meant not to be tampered, but many users did not understand that restriction which gave a lot of heavy support cases.
Right, makes sense. I’m a sys admin by trade though and I’d still like to comfortably administer my HAOS installation, and not through some GUI add-ons if possible .
Is that basically not possible then without manually copying static binaries to one of those writeable directories and other not-so-nice hacks?
Docker install doesn’t support addons and 1-click updates from the UI though, right? That’s a bummer. Supervised was really a great middle ground for folks like me I think haha.
I’ll have to get used to using GUI addons and the very limited access to the host system I guess
In reality it was not a middle ground, but people did not follow the set of restrictions.
It meant they did not run a supported installation anyway.
Lazydocker and portainer were some of the apps that caused issues, because the supervisor expect to be the sole handler of Docker containers.
I once so tan a Supervised installation, but found that running HAOS in one VM and move the rest to another VM was pretty easy and HA was sort of fully maintained in that setup.
The rest of the stuff could be handled by the other VM.
Yeah I mean I’m not going to argue that - makes a lot of sense and I’m sure the team had a lot of unnecessary headache due to this.
I see you’re active in my other recent thread too - this is an example of a technical issue somewhere that I can hardly diagnose without installing some additional tools though
I see some folks mentioned 