docker-nvidia-glx-desktop

Xfce Desktop container designed for Kubernetes supporting OpenGL GLX and Vulkan for NVIDIA GPUs with WebRTC and HTML5, providing an open source remote cloud graphics or game streaming platform. Spawns its own fully isolated X Server instead of using the host X server, therefore not requiring /tmp/.X11-unix host sockets or host configuration.

Use docker-nvidia-egl-desktop for an Xfce Desktop container which directly accesses NVIDIA (and unofficially Intel and AMD) GPUs without using an X11 Server, supports sharing a GPU with many containers, and automatically falling back to software acceleration in the absence of GPUs (but with limited graphics performance).

Read the Troubleshooting section first before raising an issue. Support is also available with the Selkies Discord.

Usage

Container startup could take some time at first launch as it automatically installs NVIDIA drivers compatible with the host.

Wine, Winetricks, Lutris, and PlayOnLinux are bundled by default. Comment out the section where it is installed within Dockerfile if the user wants to remove them from the container.

There are two web interfaces that can be chosen in this container, the first being the default selkies-gstreamer WebRTC HTML5 interface (requires a TURN server or host networking), and the second being the fallback noVNC WebSocket HTML5 interface. While the noVNC interface does not support audio forwarding and remote cursors for gaming, it can be useful for troubleshooting the selkies-gstreamer WebRTC interface or using this container with low bandwidth environments.

The noVNC interface can be enabled by setting NOVNC_ENABLE to true. When using the noVNC interface, all environment variables related to the selkies-gstreamer WebRTC interface are ignored, with the exception of BASIC_AUTH_PASSWORD. As with the selkies-gstreamer WebRTC interface, the noVNC interface password will be set to BASIC_AUTH_PASSWORD, and uses PASSWD by default if not set. The noVNC interface also additionally accepts the NOVNC_VIEWPASS environment variable, where a view only password with only the ability to observe the desktop without controlling can also be set.

The container requires host NVIDIA GPU driver versions of at least 450.80.02 and preferably 470.42.01, with the NVIDIA Container Toolkit to be also configured on the host for allocating GPUs. All Maxwell or later generation GPUs in the consumer, professional, or datacenter lineups will not have significant issues running this container, although the selkies-gstreamer high performance NVENC backend may not be available (see the next paragraph). Kepler GPUs are untested and likely does not support the NVENC backend, but can be mostly functional using fallback software acceleration.

The high performance NVENC backend for the selkies-gstreamer WebRTC interface is only supported in GPUs listed as supporting H.264 (AVCHD) under the NVENC - Encoding section of NVIDIA's Video Encode and Decode GPU Support Matrix. If your GPU is not listed as supporting H.264 (AVCHD), add the environment variable WEBRTC_ENCODER with the value x264enc, vp8enc, or vp9enc in your container configuration for falling back to software acceleration, which also has a very good performance depending on your CPU.

The username is user in both the container user account and the web authentication prompt. The environment variable PASSWD is the password of the container user account, and BASIC_AUTH_PASSWORD is the password for the HTML5 interface authentication prompt. If ENABLE_BASIC_AUTH is set to true for selkies-gstreamer (not required for noVNC) but BASIC_AUTH_PASSWORD is unspecified, the HTML5 interface password will default to PASSWD.

NOTES: Only one web browser can be connected at a time with the selkies-gstreamer WebRTC interface. If the signaling connection works, but the WebRTC connection fails, read the Using a TURN Server section.

Running with Docker

1. Run the container with Docker (or other similar container CLIs like Podman):

docker run --gpus 1 -it --tmpfs /dev/shm:rw -e TZ=UTC -e SIZEW=1920 -e SIZEH=1080 -e REFRESH=60 -e DPI=96 -e CDEPTH=24 -e VIDEO_PORT=DFP -e PASSWD=mypasswd -e WEBRTC_ENCODER=nvh264enc -e BASIC_AUTH_PASSWORD=mypasswd -p 8080:8080 ghcr.io/ehfd/nvidia-glx-desktop:latest

NOTES: The container tags available are latest and 20.04 for Ubuntu 20.04 and 18.04 for Ubuntu 18.04. Replace all instances of mypasswd with your desired password. BASIC_AUTH_PASSWORD will default to PASSWD if unspecified. The container must not be run in privileged mode.

Change WEBRTC_ENCODER to x264enc, vp8enc, or vp9enc when using the selkies-gstreamer interface if your GPU doesn't support H.264 (AVCHD) under the NVENC - Encoding section in NVIDIA's Video Encode and Decode GPU Support Matrix.

2. Connect to the web server with a browser on port 8080. You may also separately configure a reverse proxy to this port for external connectivity.

NOTES: Additional configurations and environment variables for the selkies-gstreamer WebRTC HTML5 interface are listed in lines that start with parser.add_argument within the selkies-gstreamer main script.

3. (Not Applicable for noVNC) Read carefully if the selkies-gstreamer WebRTC HTML5 interface does not connect. Choose whether to use host networking or a TURN server. The selkies-gstreamer WebRTC HTML5 interface will likely just start working if you add --network host to the above docker run command. However, this may be restricted or be undesired because of security reasons. If so, check if the container starts working after omitting --network host. If it does not work, you need a TURN server. Read the Using a TURN Server section and add the environment variables -e TURN_HOST=, -e TURN_PORT=, and pick one of -e TURN_SHARED_SECRET= or both -e TURN_USERNAME= and -e TURN_PASSWORD= environment variables to the docker run command based on your authentication method.

Running with Kubernetes

1. Create the Kubernetes Secret with your authentication password:

kubectl create secret generic my-pass --from-literal=my-pass=YOUR_PASSWORD

NOTES: Replace YOUR_PASSWORD with your desired password, and change the name my-pass to your preferred name of the Kubernetes secret with the xgl.yml file changed accordingly as well. It is possible to skip the first step and directly provide the password with value: in xgl.yml, but this exposes the password in plain text.

2. Create the pod after editing the xgl.yml file to your needs, explanations are available in the file:

kubectl create -f xgl.yml

NOTES: The container tags available are latest and 20.04 for Ubuntu 20.04 and 18.04 for Ubuntu 18.04. BASIC_AUTH_PASSWORD will default to PASSWD if unspecified.

Change WEBRTC_ENCODER to x264enc, vp8enc, or vp9enc when using the selkies-gstreamer WebRTC interface if your GPU doesn't support H.264 (AVCHD) under the NVENC - Encoding section in NVIDIA's Video Encode and Decode GPU Support Matrix.

3. Connect to the web server spawned at port 8080. You may configure the ingress endpoint or reverse proxy that your Kubernetes cluster provides to this port for external connectivity.

NOTES: Additional configurations and environment variables for the selkies-gstreamer WebRTC HTML5 interface are listed in lines that start with parser.add_argument within the selkies-gstreamer main script.

4. (Not Applicable for noVNC) Read carefully if the selkies-gstreamer WebRTC HTML5 interface does not connect. Choose whether to use host networking or a TURN server. The selkies-gstreamer WebRTC HTML5 interface will likely just start working if you uncomment hostNetwork: true in xgl.yml. However, this may be restricted or be undesired because of security reasons. If so, check if the container starts working after commenting out hostNetwork: true. If it does not work, you need a TURN server. Read the Using a TURN Server section and fill in the environment variables TURN_HOST and TURN_PORT, then pick one of TURN_SHARED_SECRET or both TURN_USERNAME and TURN_PASSWORD environment variables based on your authentication method.

Using a TURN Server

Note that this section is only required for the selkies-gstreamer WebRTC HTML5 interface. For an easy fix to when the signaling connection works, but the WebRTC connection fails, add the option --network host to your Docker command, or uncomment hostNetwork: true in your xgl.yml file when using Kubernetes (note that your cluster may have not allowed this, resulting in an error). This exposes your container to the host network, which disables network isolation. If this does not fix the connection issue (normally when the host is behind another firewall) or you cannot use this fix for security or technical reasons, read the below text.

In most cases when either of your server or client has a permissive firewall, the default Google STUN server configuration will work without additional configuration. However, when connecting from networks that cannot be traversed with STUN, a TURN server is required. Provide the TURN server address, port, and shared secret in order to take advantage of the TURN relay capabilities and improve connection success.

Open Relay is a free TURN server instance that may be used for personal purposes, but may not be optimal for production usage.

An open source TURN server that can be used is coTURN, and an example container coturn/coturn:latest is available. For dynamic IP addresses, dynamic-coturn is a container implementation which restarts the TURN server whenever the public IP address gets changed. Pion TURN is another TURN server implementation compatible with all major operating systems including Windows.

It is possible to install coTURN on your own server or PC, as long as ports can be opened. In short, /etc/turnserver.conf must have either use-auth-secret and static-auth-secret=(PUT RANDOM 64 BYTE BASE64 KEY HERE), or lt-cred-mech and user=yourusername:yourpassword. Other optional, but useful parameters include min-port= and max-port= for setting your relay ports between TURN servers, and cert= and pkey= options for TURN over TLS/DTLS. Install coTURN from your package repository, or use its container image with Docker/Podman or Kubernetes.

Deploy coTURN with Docker

In order to deploy a coTURN container, use the following command (consult this example configuration for more options which may also be used as command line arguments). You should be able to expose these ports to the internet. Change -p 49160-49200:49160-49200/udp and --min-port=49160 --max-port=49200 as appropriate. Simply using --network host instead of specifying -p 49160-49200:49160-49200/udp is also fine if possible.

For time-limited shared secret TURN authentication:

docker run -d -p 3478:3478 -p 3478:3478/udp -p 49160-49200:49160-49200/udp coturn/coturn -n --min-port=49160 --max-port=49200 --use-auth-secret --static-auth-secret=(PUT RANDOM 64 BYTE BASE64 KEY HERE)

For legacy long-term TURN authentication:

docker run -d -p 3478:3478 -p 3478:3478/udp -p 49160-49200:49160-49200/udp coturn/coturn -n --min-port=49160 --max-port=49200 --lt-cred-mech --user=yourusername:yourpassword

If you want to use TURN over TLS/DTLS, you must have a valid hostname, and also provision a valid certificate issued from a legitimate certificate authority such as ZeroSSL (Let's Encrypt may have issues depending on the OS) and provide the certificate and private files to the coTURN container with -v /mylocalpath/coturncert.pem:/etc/coturncert.pem -v /mylocalpath/coturnkey.pem:/etc/coturnkey.pem and add command line arguments -n --cert=/etc/coturncert.pem --pkey=/etc/coturnkey.pem.

More information available in the coTURN container image or the coTURN repository website.

Configuring with Docker

With Docker (or Podman), use the -e option to add the TURN_HOST, TURN_PORT environment variables. This is the hostname or IP and the port of the TURN server (3478 in most cases).

You may set TURN_PROTOCOL to tcp if you are only able to open TCP ports for the coTURN container to the internet, or if the UDP protocol is blocked or throttled in your client network. You may also set TURN_TLS to true with the -e option if TURN over TLS/DTLS was properly configured.

You also require to provide either just TURN_SHARED_SECRET for time-limited shared secret TURN authentication, or both TURN_USERNAME and TURN_PASSWORD for legacy long-term TURN authentication, depending on your TURN server configuration. Provide just one of these authentication methods, not both.

Deploy coTURN With Kubernetes

You are recommended to use a ConfigMap for creating the configuration file for coTURN. Use the example coTURN configuration as a reference to create a ConfigMap which mounts to /etc/turnserver.conf. The only mandatory lines are either use-auth-secret and static-auth-secret=(PUT RANDOM 64 BYTE BASE64 KEY HERE) or lt-cred-mech and user=yourusername:yourpassword.

Use Deployment or DaemonSet and use containerPort and hostPort under ports: to open port 3478 (or any other port you set in /etc/turnserver.conf with listening-port=).

Then you must also open all ports between min-port= and max-port= that you set in /etc/turnserver.conf, but this may be skipped if hostNetwork: true is used instead.

Under args: set -c /etc/turnserver.conf and use the coturn/coturn:latest image.

If you want to use TURN over TLS/DTLS, use cert-manager to issue a valid certificate with the correct hostname from preferably ZeroSSL (Let's Encrypt may have issues based on the OS), then mount the certificate and private key in the container. Do not forget to include the options cert= and pkey= in /etc/turnserver.conf to the correct path of the certificate and the key.

More information available in the coTURN container image or the coTURN repository website.

Configuring With Kubernetes

Your TURN server will use only one out of two ways to authenticate the client, so only provide one type of authentication method. The time-limited shared secret TURN authentication requires to only provide the Base64 encoded TURN_SHARED_SECRET. The legacy long-term TURN authentication requires to provide both TURN_USERNAME and TURN_PASSWORD credentials.

Time-Limited Shared Secret Authentication

1. Create a secret containing the TURN shared secret:

kubectl create secret generic turn-shared-secret --from-literal=turn-shared-secret=MY_TURN_SHARED_SECRET

NOTES: Replace MY_TURN_SHARED_SECRET with the shared secret of the TURN server, then changing the name turn-shared-secret to your preferred name of the Kubernetes secret, with the xgl.yml file also being changed accordingly.

2. Uncomment the lines in the xgl.yml file related to TURN server usage, updating the TURN_HOST and TURN_PORT environment variable as needed:

- name: TURN_HOST
  value: "turn.example.com"
- name: TURN_PORT
  value: "3478"
- name: TURN_SHARED_SECRET
  valueFrom:
    secretKeyRef:
      name: turn-shared-secret
      key: turn-shared-secret
- name: TURN_PROTOCOL
  value: "udp"
- name: TURN_TLS
  value: "false"

NOTES: It is possible to skip the first step and directly provide the shared secret with value:, but this exposes the shared secret in plain text. Set TURN_PROTOCOL to tcp if you were able to only open TCP ports while creating your own coTURN Deployment/DaemonSet, or if your client network throttles or blocks the UDP protocol.

Legacy Long-Term Authentication

1. Create a secret containing the TURN password:

kubectl create secret generic turn-password --from-literal=turn-password=MY_TURN_PASSWORD

NOTES: Replace MY_TURN_PASSWORD with the password of the TURN server, then changing the name turn-password to your preferred name of the Kubernetes secret, with the xgl.yml file also being changed accordingly.

2. Uncomment the lines in the xgl.yml file related to TURN server usage, updating the TURN_HOST, TURN_PORT, and TURN_USERNAME environment variable as needed:

- name: TURN_HOST
  value: "turn.example.com"
- name: TURN_PORT
  value: "3478"
- name: TURN_USERNAME
  value: "username"
- name: TURN_PASSWORD
  valueFrom:
    secretKeyRef:
      name: turn-password
      key: turn-password
- name: TURN_PROTOCOL
  value: "udp"
- name: TURN_TLS
  value: "false"

NOTES: It is possible to skip the first step and directly provide the TURN password with value:, but this exposes the TURN password in plain text. Set TURN_PROTOCOL to tcp if you were able to only open TCP ports while creating your own coTURN Deployment/DaemonSet, or if your client network throttles or blocks the UDP protocol.

Comparison

docker-nvidia-egl-desktop: It's generally recommended to use docker-nvidia-glx-desktop when possible for maximum capabilities and performance. It starts its own X server inside the container without exposure to security risks. However, docker-nvidia-egl-desktop is versatile in various environments and has less processes running, meaning less possible errors. It is also possible to be used in HPC clusters with Apptainer/Singularity available, and sharing a GPU with multiple containers is also possible. Unofficial support for Intel and AMD GPUs is also available.

Sunshine: This repository is an open-source server for NVIDIA's GameStream protocol, supporting all clients that can install Moonlight. Try it if you don't need username/password authentication and you don't need to use containers. Games on Whales is a container implementation of Sunshine. However, many container ports have to be accessible to the internet, and because of its requirement for the /dev/uinput device, unsafe privileged access for containers are required. The selkies-gstreamer project, which is integrated to our container, does not require more than one port open from the container (TURN server may be required but can be deployed in a different environment with flexibility), and has almost equal performance while using only a web browser as a client.

x11docker: This has a lot of features and is very solid if you are the sole user in full control of the host. However, it starts a lot of processes in the host and is nearly impossible to contain the environment. Kubernetes is also not supported. The docker-nvidia-glx-desktop repository contains everything in the container, with the only requirement being the NVIDIA Container Toolkit with adequate NVIDIA_DRIVER_CAPABILITIES, meaning that the container is portable anywhere Docker/Podman or Kubernetes can be run.

Xpra: This is a feature-complete all-in-one remote desktop application optimized for Linux, although not exactly meant for full screen workloads and its HTML5 web interface is not optimized for intensive graphics workloads. Supports various protocols and various hardware acceleration methods.

KasmVNC: This almost landed as a replacement for the existing noVNC fallback installation, as it incorporates improved functionalities. However, performance compared to x11vnc combined with noVNC was not much better.

Parsec: Parsec is not open-source. However, it brings top-level performance on Windows or Mac hosts. Try it if you don't need to use containers. But the selkies-gstreamer project uses the same APIs and isn't that far back in terms of performance.

CloudRetro and CloudMorph: This uses WebRTC in a web browser, like the selkies-gstreamer project. The principles of this project are pretty similar to our project. However, hardware acceleration across various GPUs is currently not implemented. Hardware acceleration is critical to remote desktop and workload performance, and therefore you should use our repository if you need hardware acceleration.

neko: Uses WebRTC in a web browser with a text chat, containers can be used, but hardware acceleration is very limited. Use this if you want the best performance while requiring multiple users to be able to access the screen. However, note that you can always use conference software such as Zoom, Jitsi, or BigBlueButton to share your screen while using our container.

RustDesk: This is an open-source TeamViewer or AnyDesk. You can use this to have other people control your container if you need to.

Weylus: This is a very interesting project, and has many technologies in common. Use this if you want to turn your tablet or smartphone to a graphic tablet for your PC.

GamingAnywhere: This is the father of all open-source remote desktop and game streaming protocols. However, it has been created a long time ago and thus reached its end of life.

Troubleshooting

The container doesn't work.

Check that the NVIDIA Container Toolkit is configured in the host. If you did that, scroll down.

The container doesn't work if an existing GUI or X server is running in the host outside the container. / I want to use --privileged mode or --cap-add to my containers.

In order to use an X server on the host for your monitor with one GPU, and then provision other GPUs for the containers, you must change your /etc/X11/xorg.conf configurations.

First, use sudo nvidia-xconfig --no-probe-all-gpus --busid=$BUS_ID --only-one-x-screen to generate /etc/X11/xorg.conf where BUS_ID is generated with the below script. Set GPU_SELECT to the ID of the specific GPU you want to provision from nvidia-smi.

HEX_ID=$(nvidia-smi --query-gpu=pci.bus_id --id="$GPU_SELECT" --format=csv | sed -n 2p)
IFS=":." ARR_ID=($HEX_ID)
unset IFS
BUS_ID=PCI:$((16#${ARR_ID[1]})):$((16#${ARR_ID[2]})):$((16#${ARR_ID[3]}))

Then, edit /etc/X11/xorg.conf and add the following to the end. If you want to use the containers in privileged mode, add this section to the /etc/X11/xorg.conf file of all containers as well.

Section "ServerFlags"
    Option "AutoAddGPU" "false"
EndSection

Reference

If you restart your OS or the Xorg server, you will now be able to use one GPU for your host X server and your real monitor, and use the rest of the GPUs for the containers.

Then, you must avoid the GPU of which you are using for your host X server. Use docker --gpus '"device=1,2"' to provision GPUs with device IDs 1 and 2 to the container, avoiding the GPU with the ID of 0 that is used by the host X server, if you set GPU_SELECT to the ID of 0. Note that --gpus 1 means any single GPU, not the GPU device ID of 1.

Vulkan does not work.

Make sure that the NVIDIA_DRIVER_CAPABILITIES environment variable is set to all or includes all of utility, graphics, video, and display. The display capability is especially crucial to Vulkan, but the container does start without display, even with its name.

The container doesn't work if I set the resolution above 1920 x 1200 or 2560 x 1600 in 60 hz.

Short Answer

If your GPU is a consumer or professional GPU, change the VIDEO_PORT environment variable from DFP to DP-0 if DP-0 is empty, or any empty DP-* port. Set VIDEO_PORT to where your monitor is connected if you want to show the remote desktop in a real monitor. If your GPU is a Datacenter (Tesla) GPU, keep the VIDEO_PORT environment variable to DFP, and your maximum resolution is at 2560 x 1600. To go above this restriction, you may set VIDEO_PORT to none, but you must use borderless window instead of fullscreen, and this may lead to quite a lot of applications not starting, showing errors related to XRANDR or RANDR.

Long Answer

The container simulates the GPU being virtually plugged into a physical DVI-D/HDMI/DisplayPort digital video interface in consumer and professional GPUs with the ConnectedMonitor NVIDIA driver option. The container uses virtualized DVI-D ports for this purpose in Datacenter (Tesla) GPUs. The ports to be used should only be connected with an actual monitor when the user wants the remote desktop screen to be shown on that monitor. If you want to show the remote desktop screen spawned by the container in a physical monitor, connect the monitor and set VIDEO_PORT to the the video interface identifier that is connected to the monitor. Manually specify a video interface identifier that is not connected to a monitor in VIDEO_PORT if you have a physical monitor connected and want to show the screen to the monitor. VIDEO_PORT identifiers and their connection states can be obtained by typing xrandr -q when the DISPLAY environment variable is set to the number of the spawned X server display (for example :0). As an alternative, you may set VIDEO_PORT to none (which effectively sets --use-display-device=None), but you must use borderless window instead of fullscreen, and this may lead to quite a lot of applications not starting because the RANDR extension is not available in the X server.

NOTES: Do not start two or more X servers for a single GPU. Use a separate GPU (or use Xvfb/Xdummy/XVnc without hardware acceleration to use no GPUs) if you need a host X server unaffiliated with containers, and do not make the GPU available to the container runtime.

Since this container simulates the GPU being virtually plugged into a physical monitor while it actually does not, make sure the resolutions specified with the environment variables SIZEW and SIZEH are within the maximum size supported by the GPU. The environment variable VIDEO_PORT can override which video port is used (defaults to DFP, the first interface detected in the driver). Therefore, specifying VIDEO_PORT to an unplugged DisplayPort (for example numbered like DP-0, DP-1, and so on) is recommended for resolutions above 1920 x 1200 at 60 hz, because some driver restrictions are applied when the default is set to an unplugged physical DVI-D or HDMI port. The maximum size that should work in all cases is 1920 x 1200 at 60 hz, mainly for when the default VIDEO_PORT identifier DFP is not set to DisplayPort. The screen sizes over 1920 x 1200 at 60 hz but under the maximum supported display size specified for each port (supported by GPU specifications) will be possible if the port is set to DisplayPort (both physically connected or disconnected), or when a physical monitor or dummy plug to any other type of display ports (including DVI-D and HDMI) has been physically connected. If all GPUs in the cluster have at least one DisplayPort and they are not physically connected to any monitors, simply setting VIDEO_PORT to DP-0 is recommended (but this is not set as default because of legacy GPU compatibility reasons).

Datacenter (Tesla) GPUs seem to only support resolutions of up to around 2560 x 1600 at 60 hz (VIDEO_PORT must be kept to DFP instead of changing to DP-0 or other DisplayPort identifiers). The K40 (Kepler) GPU did not support RandR (required for some graphical applications using SDL and other graphical frameworks). Other Kepler generation Datacenter GPUs (maybe except the GRID K1 and K2 GPUs with vGPU capabilities) are also unlikely to support RandR, thus Datacenter GPU RandR support probably starts from Maxwell. Other tested Datacenter GPUs (V100, T4, A40, A100) support all graphical applications that consumer GPUs support. However, the performances were not better than consumer GPUs that usually cost a fraction of Datacenter GPUs, and the maximum supported resolutions were even lower.

This project involved a collaboration effort with Itopia's Dan Isla (founder of the Selkies Project), incorporating the selkies-gstreamer WebRTC remote desktop streaming application. Commercial support for this container may be available from Itopia.

This work was supported in part by NSF awards CNS-1730158, ACI-1540112, ACI-1541349, OAC-1826967, OAC-2112167, CNS-2120019, the University of California Office of the President, and the University of California San Diego's California Institute for Telecommunications and Information Technology/Qualcomm Institute. Thanks to CENIC for the 100Gbps networks.