WTH Do I have to do so many restarts to get my system up to date?

It would be really handy to reduce the number of restarts. Ideally when I’m ready to update everything, I could apply a Home Assistant update, a Home Assistant OS update, and addon updates (like Zwave JS UI, Zigbee2MQTT, Mosquitto) and THEN tell it to reboot once instead of after each update.

(Yes, I know technically Home Assistant don’t have to be restarted to update addons like MQTT, Zwave, or Zigbee, but those addon restarts can be disruptive enough, making large number of things go unavailable, that it would be nice to defer the restart until I reboot.)

Previous very verbose WTH: Reduce restarts during maintenance: Update HaOS, Home Assistant, Add-ons but defer restarts until reboot

You do not have to restart Home Assistant for add-on updates. Or for supervisor updates.

You can update all your third party integrations and Home Assistant then only restart once.

An OS update does take one additional restart.

So that is two restarts.

I live on the edge and update all add ons and hacs stuff, then supervisor and then do the main Hass OS one last, which forces a reboot.

Haven’t had an issue this far and it only takes 1 reboot.

2 Likes

I said the following which I guess wasn’t as clear as I could have made it, maybe I should have said downtime instead of restarts?

If I have to update something like Zwave JS UI, Z2M, Mosquitto, when the addon restarts lots of devices go unavailable, running automations break.

If I have updates for addons, home assistant, and home assistant OS, there are 3 times when there will be some downtime where running automation will break, events may be missed, etc.

I know that I could use docker commands to change the tags for the various containers and not restart anything until I reboot or restart the containers.

AFAICT, ha core update and ha addon update do not have an option to update the containers will out restarting them.

Either way, I thought it would be reasonable for the UI to have an option for updating without restarting so that all the downtime could be combined into one reboot/restart.

Even better would be a UI that let me check all of the things I want to update and then hit “GO” once.

No, that’s my fault. I skimmed it. Sorry.

Using Zigbee2MQTT (and Mosquitto) all my entities recover within 10 seconds of either add-on restart. I guess it could interrupt running automations but none of my automations wait/delay so they execute quickly. And my triggers are set up to ignore to: unavailable/unknown. Some will trigger when recovering from: unknown/unavailable to: something but that is exactly what I want.

Unfortunately not possible. Even Windows applications have to shut-down to update.

That is possible for 3rd party integrations and core you can do all those at once. Just not add-ons at the moment. Requesting this for add-ons is a good WTH.