Skip to content

Latest commit

 

History

History
102 lines (81 loc) · 3.9 KB

README.md

File metadata and controls

102 lines (81 loc) · 3.9 KB
Raw

AndroidTV adb-server

Ever since upgrading my NVIDIA Shield TV to Android 11, it started behaving less well in combination with Home Assistant's AndroidTV integration.

In essence it works fine, but while switched on the player's state randomly jumps to standby for 10-second intervals every couple of minutes. This behaviour can be worked around in most automations, but the fact of the matter remains the player's state is often incorrect...

Anecdotal evidence suggests that running a "native" adb-server (instead of the Python implementation provided with Home Assistant) might resolve the issue. As such, this Docker-container tools up an adb-server specifically for that purpose.

Further investigation shows it's most probably "Shield Experience 9.1"-related (aka, the Android 11 upgrade), with a likely fix incoming via python-androidtv/#329.

Anyway...

adb-server

The principle behind the container is simple: It starts an adb-server and subsequently attempts to adb connect to a given AndroidTV-device every 30-seconds (connect is a no-op if already connected). This ensures there's always an active ADB-connection to the AndroidTV.

The Home Assistant integration uses its Python-client to communicate with the already/always connected "native" adb-instance.

An additional benefit of holding the ADB-connection "outside" of Home Assistant is, that it's now possible for other clients to utilise it too — Home Assistant doesn't hold an exclusive lock on the connection anymore...

As such, there might be some merit to keeping ADB running this way; regardless of what originally prompted the approach.

"Remote" connections

To "remotely" connect a client to the adb-server running in the container, do the following:

export ANDROID_ADB_SERVER_ADDRESS=127.0.0.1
adb devices # Should show a connected AndroidTV device

This works from the machine running the container. Note that it is required to use 127.0.0.1. Using localhost will cause adb to automatically spawn a server (defeating the purpose).

Actual remote connections are also possible. In that case, ensure the container's port-forwarding (see below) is setup as 5037:5037/tcp (without the leading 127.0.0.: as that limits it to local connections only). Note that configuration access in this way poses a substantial security-risk to your Android device...

Setup

Use the below 📄 docker-compose.yml (start using docker-compose up -d):

version: "2.3"
services:
  androidtv:
    image: androidtv
    # Point to the location of the Dockerfile
    build: ./extras/androidtv
    restart: unless-stopped
    network_mode: bridge
    ports:
      # Remove "127.0.0.1:" if you intended to access adb remotely
      - 127.0.0.1:5037:5037/tcp
    volumes:
      # This folder should contain "adbkey" and "adbkey.pub"; if omitted, new
      # keys are generated every time the container is created
      - ./extras/androidtv/keys:/androidtv/keys
    environment:
      # Hostname or IP address of your AndroidTV device; port 5555 is assumed
      - ANDROIDTV_DEVICE=nvidia-shield-tv
      # Optional: Enable verbose output; normally only "adb" messages are
      # logged — <https://developer.android.com/studio/command-line/variables>
      - ADB_TRACE=all
      # Optional: Drop permissions to the provided UID/GID-combination
      - PUID=1000
      - PGID=1000

Keys

Home Assistant's AndroidTV integration by default keeps its keys in 📂 .storage (named 📄 androidtv_adbkey and 📄 androidtv_adbkey.pub). Rename those to 📄 adbkey and 📄 adbkey.pub respectively and feed them into the container (see above) to prevent having to reauthenticate with your AndroidTV-device.