Routine updates should not leave users spending hours recovering their smart homes

Configuration Zigbee
1

Today I installed a routine Home Assistant update.

I expected bug fixes and improvements.

Instead, I spent my entire Saturday evening trying to recover my smart home.

After the update:

  • Zigbee2MQTT stopped controlling my Aqara devices.
  • Devices were still visible.
  • MQTT was connected.
  • Telemetry was still received.
  • Commands failed with MAC_CHANNEL_ACCESS_FAILURE.
  • While trying to recover the network, I even lost one device and could no longer pair it again.

The most frustrating part is not even the bug itself.

It is the user experience.

As an end user, I have no practical way to know whether the problem is Home Assistant Core, Zigbee2MQTT, zigbee-herdsman or another dependency.

From my perspective, only one thing happened:

I installed an update offered by Home Assistant.

After installing it, my previously working smart home stopped working.

I understand that software has bugs.

What I do not understand is why routine updates can leave users spending hours debugging their homes instead of using them.

There should be a simple rollback option that restores yesterday's working system within minutes.

Instead, I spent hours checking logs, MQTT, Zigbee2MQTT, coordinator settings, channels and network configuration without finding a solution.

This is not the experience people expect from software that controls real devices like lights, plugs and ventilation.

Please make update reliability a higher priority.

A smart home should survive routine updates.

If an update does break something, users should be able to return to the previous working version quickly.

Today I wanted to relax after work.

Instead, I spent my entire evening trying to recover from a Home Assistant update.

That should never become the normal Home Assistant experience.

Is improving update reliability and providing an easy rollback path something the Home Assistant project is planning to address?

2

This is due to interference in the zigbee spectrum. Nothing to do,with the update. It took me one internet search and a quick read of the top result (on the z2m github) to find that out.

Additionally updating the core has little effect on the Z2M container.

There is. In fact there are two.

  1. Restore from backup. Replaces everything as it was when the backup was made. The last time I had to do this was when I moved to a new NUC many years ago.

Or

  1. Use the HA CLI to rollback to the previous version. Only replaces the core version. On the rare occasion when I have an issue with testing the beta version of HA I just do this. You can also use this for issues with the HAOS (different command).