It will be improved with Marketplace replacing HACS next month. No need to restart after installing a new third party integration. You still need to restart after updating them though. It’s just the way HA is built. The Devs are aware but it is a massive job to change the entire core architecture.
HA always kind of reminds me of the old days with Windows NT; the standard joke was “You have moved the mouse; you must restart.” If you made almost ANY sort of config change, you’d need a reboot.
It reminds me of blue-screen-of dead
At a linux course i made a prank with a linux “Guru” he hated and banned Windows users as amateurs , and was not “familiar” with nt4(which were our usual study pc’s)
@ startup, open browser>full-screen, with url to a simple webpage
No-mouse-click and a background with a Blue-Screen-of Dead pic … He kept clicking and Rebooting and got more and more frustrated
He figured eventually out that i couldn’t hold me from laughing, and teasing him with small Hints/jokes about linux users, until i told him to hit the windows button, or “open” the task-manger , the whole class was dying … he wasn’t so cocky after that
I don’t think I was imagining it when I had two 3rd party integrations that needed an update at the same time and I needed to do a restart for each one. I haven’t had the opportunity to repeat this experiment yet, though . . .
It would be great if all Integrations were contained inside an App (like the Matter Server) so the core HA state machine etc. could stay up while the Integrations are updated and restarted.
Like many other Systems, there is nothing that prevent you from installing both Updates, and Then Restart 1 time
As 3rd Integrations is capable of installing i.e dependencies which can cause conflicts in CORE as well as Other Integrations, Your 3rd part is “responsible” either way
And if an Integration i.e downgrade/upgrade a library, which initially doesn’t seems to causes any “breaking changes” , You most like don’t “notice” anything, before next CORE up date, with a “Bump”, causing your integration to i.e crash … And thus You blame the Core Update , either way
However CORE Dependencies have Priority , so you 3rd part “Upgrade” most likely fails or causes direct “noticeable” Errors
NOTE: i’ve seen many which have installed bunch of integrations, and even after years of using HA, have no freaking ideas where to find CORE-RAW log
They install, and if they initially find that things seamlessly works, they will “sleep tight” unaware the Logger is spewing, usage increasing, and a crash is around the corner
PS: it’s also an answer to the WTH, Integrations Do install in CORE also
That’s what I’ve always done. If there is a HASS update/Supervisor update/HAOS update then, I just do those last and it’ll do it all at one.
With that said, they might be thinking of exactly that … I don’t think you can queue a HASS update and a Supervisor/HAOS update to happen at the same time, you’ve got to do them individually.
I Don’t know, it’s not quite clear from anyone above
For Good reasons i’ve always Separated HACS Update from i.e OS/CORE, As i have an interest in seeing/Knowing whether CORE/OS creates errors in the Logs, as well as whether HACS updates create Errors, Any way else would be stupid.
In Regards to Supervisor, i don’t remember last time i ever had to do anything to Upgrade that, as it’s usually an Auto-Update, at random time
And what you mean by an HASS update, i have no idea, HA-OS/CORE is optional ( Most of the time ) if not specificly “Noted”