I installed hass.io using this script, very successfully.
I do find that HA does become unstable after a few days so I’d like to set up a regular restart of the whole system.
What is the safest way?
Restart the supervisor? systemctl restart hassio-supervisor.service
Or another way?
I found if you install HassOs, at least the 64-bit VM, it gets unstable & need a reboot. I have not seen that issue with Hassio installed on Raspian on my RPi 3B+
In my personal opinion as a former Unix administrator, I think HassOs is likely unstable.
Hassio is great to get started but I may move away to a Home Assistant installation to gain more flexibility.
This is a docker install on top of a DietPi base running in a VM on a micro server. The DietPi base is extremely stable so I’m sure that it is Home Assistant / HassIO that is the issue (history just stops updating for instance).
A restart, and all is well again, but I find that just starting it from within HassIO is not sufficient and it really needs the whole docker environment restarted. It is about how best to do that
Fair enough. I have only run my docker version of Hassio for less than a week.
Something is wrong. You should NOT have to do that.
Maybe so, but right now I do, so it would be nice to know what the recommended way is to restart the system (does not need a reboot).
However, I have just discovered the disk was full so that may be the problem! Fortunately it was on a VM and running DietPi means there is a small utility to expand the FS so all good now.
Could really do with a low disk space alert.
You can create your own disk space alert.
https://www.home-assistant.io/components/sensor.systemmonitor/
I can’t agree with this. I know 32bit is recommended but I ran it for months with no compatability issues… and that was on an early version just after released. Apparently there are some addons that don’t work but I never found one.
There’s nothing wrong with HassOS in either 32 or 64 bit.
Well that will do it for sure.
Yes but it would be much better when, on hitting ‘update’, it came back with an ‘insuffcient space’ message rather than just returning you to the ‘please update’ with no explanation as to why it failed. Plenty of other builtin alerts so why not one for what is a pretty critical issue???
And if it is so easy to do…
[edit] Still hoping for an answer to the original question.
There are services you can use in automations. There’s homeassistant.restart (only restarts HA) or hassio.host_reboot (reboots the whole device/vm/etc) which you can find in the services dev tool dropdown. Restarting the supervisor only restarts the supervisor itself (which might be useful for checking for updates).