In what area(s)?

/area runtime

What version of Dapr?

CLI version: 1.15.x
Runtime version: 1.15.x

Expected Behavior

I should be able to use dapr runtime in standalone mode from a virualized machine running in the host machine where DAPR is running.

Environment:

- Host (MacOS)
  ├── Docker
  │   └── DAPR runtime
  └── Parallels
      └── VM (Windows)
          └── App using DAPR runtime running on the Host

In the Host machine the DAPR runtime has been initialized with the command

dapr init

In the VM, DAPR has been initialized with the command

dapr init --slim

and the resource files have been updated replacing "localhost" with the IP of the Host.

Actual Behavior

With DAPR 1.15, the scheduler is reached from the VM , but then it broadcast its address overriding the IP address given in the arguments of the DAPR sidecar, so it broadcast its address again with "localhost" and the App in the VM is not able anymore to reach the scheduler.

This is mainly because of the PR #1480 that force the overriding in standalone mode.
I understand that the new flag has been created for fixing the issue dapr/dapr#8412, but it has broken my scenario.

The override should be intentional and not forced, or at least there should be a way out. Because otherwise I don't understand the meaning of running the sidecar passing the IP address of the scheduler. It is useful to have the possibility to pass the IP of the scheduler and placement services when you run the sidecar, but these setting should not be overridden.

I have tried to recreate the scheduler container manually, without the argument --override-broadcast-host-port=localhost:50006 and it seems that it fixes the initialization of the scheduler, but there are still other issues that make my scenario not working.

In the end I had to roll back to the runtime 1.14.4.