Was seeing something a bit fishy with one of the entities in Home Assistant so I hit the old "Reload" on the integration.
Now instead of being fishy it's unavailable.
Confirmed it was still reachable on the network.
Reload again. More items are unavailable.
Reload again. Even more items are unavailable, and those that were updating before are now stuck at their values.
Do I keep reloading until everything is unavailable or do I give up?
I have used "qemu-img" to expand the qcow2 volume on which Home Assistant is storing its information.
Problem solved!
Also Home Assistant doesn't like my setup on a qemu virt and would prefer to be on bare metal or in Docker.
Searching for the error message from ESPhome shows that other people have the same problem as me and there's no clear error or fix: when ESPhome first downloads parts of the toolchain it can fail and not properly be detected (well, in spite of the page of red errors) and even if you delete the cache it still doesn't fix it as the toolchain lives outside the cache.
One fix I found buried in an archived comment is to remove and add the entire Add-on. Fixed it for me!
Future Sprog