rpi_websocket_server

Web user interface for Raspberry Pi, available over the Internet and using websockets for bi-directional low latency communication..

install Docker on Raspberry Pi

https://withblue.ink/2020/06/24/docker-and-docker-compose-on-raspberry-pi-os.html

run with docker-compose

sudo docker-compose up

run with docker run

build the Docker image

[optional 1] Sometimes it's good to remove old "dangling images" at first (add flag -a to remove all images):

sudo docker image prune

[optional 1] Sometimes it's good also to remove old containers:

sudo docker container prune

Run:

docker build -t hnet_ngnix_srv .

run command for production

sudo docker run --rm --privileged -it \
--env HOSTNAME='dckrtest' \
--env JOINCODE='fc94:b01d:1803:8dd8:3333:2222:1234:1111/xxxxxxxxxxxxxxxxx' \
--env BUTTON_PIN='26' \
--env LED_PIN='4' \
-v dckrtest_v:/var/lib/husarnet \
hnet_ngnix_srv

if you place an .env file with environmental variables listed like:

HOSTNAME=dckrtest
JOINCODE=fc94:b01d:1803:8dd8:3333:2222:1234:1111/xxxxxxxxxxxxxxxxx
BUTTON_PIN=26
LED_PIN=4

You can then run the container like this:

sudo docker run --rm --privileged -it \
--env-file ./.env \
-v dckrtest_v:/var/lib/husarnet \
hnet_ngnix_srv

TODO: --privileged flag is only temporary to access RaspberryPi GPIO. Other options are mentioned here: https://stackoverflow.com/questions/30059784/docker-access-to-raspberry-pi-gpio-pins or --cap-add SYS_RAWIO option. But neither works ...

run command for development

If you run a container like this, and make changes in files on hosts, the changes will be available in the container file system

sudo docker run --rm --privileged -it \
--env HOSTNAME='dckrtest' \
--env JOINCODE='fc94:b01d:1803:8dd8:3333:2222:1234:1111/xxxxxxxxxxxxxxxxx' \
--env BUTTON_PIN='26' \
--env LED_PIN='4' \
-v dckrtest_v:/var/lib/husarnet \
-v "/home/pi/tech/rpi_websocket_server/backend_src:/app:ro" \
-v "/home/pi/tech/rpi_websocket_server/frontend_src:/var/www/html:ro" \
hnet_ngnix_srv

or (if you placed environment variables in .env file):

sudo docker run --rm --privileged -it \
--env-file ./.env \
-v dckrtest_v:/var/lib/husarnet \
-v "/home/pi/tech/rpi_websocket_server/backend_src:/app:ro" \
-v "/home/pi/tech/rpi_websocket_server/frontend_src:/var/www/html:ro" \
hnet_ngnix_srv

open a website hosted by container

On any other computer in the same Husarnet network as the container, open a Firefox and open: http://mycontainer:80 URL.

INFO

Volumes (named)

It's good for:

store container data (even -rm don't remove that)
share data between containers

Add -v dckrhostname_v:/var/lib/husarnet as a volume to retain Husarnet Client data if we remove the container. dckrhostname_v could be whatever. You can name it the same as container, however multiple container can still access it.

Bind mounts

It's good only for:

easy access to container logs
developing a code on your host that can be executed on container without re-running it
use it only in development, not in production (this is why you still need to COPY project files in the Dockerfile)

Add -v "/home/pi/tech/rpi_websocket_server/temp:/app/test:ro as a volume to have a share folder between my host and container (add another annonymous volume -v /app/test/whatever (anonymous volume) to exclude whatever folder from container file system from synchronization). You can no override files created by container in this shared folder from a host, but the container can override files created by host. ro means read-only, so container can not write here.

DominikN/rpi_websocket_server