"Create React App" with Docker - reactjs

I was wondering if anyone had any experience using create-react-app with docker. I was able to get it set up with a Dockerfile like:
from node
RUN mkdir /src
WORKDIR /src
ADD package.json /src/package.json
RUN npm install
EXPOSE 3000
CMD [ "npm", "start" ]
And then used a docker-compose file like:
app:
volumes:
- "./app:/src"
ports:
- "3000:3000"
- "35729:35729"
build: ./app
This allowed me to start up the container and view the app. However livereload didn't work when saving files in the mounted volume and webpack created several .json.gzip files in the src directory.
Any suggestions for getting this working correctly?

Yeah, as aholbreich mentioned, I'd use npm install / npm start locally on my machine for development, just because it's so easy. It's probably possible with docker-compose, mounting volumes etc. too, but I think it could be a bit fiddly to set up.
For deployment you can then very easily use a Dockerfile. Here's an example Dockerfile I'm using:
FROM node:6.9
# Create app directory
RUN mkdir -p /src/app
WORKDIR /src/app
# to make npm test run only once non-interactively
ENV CI=true
# Install app dependencies
COPY package.json /src/app/
RUN npm install && \
npm install -g pushstate-server
# Bundle app source
COPY . /src/app
# Build and optimize react app
RUN npm run build
EXPOSE 9000
# defined in package.json
CMD [ "npm", "run", "start:prod" ]
You need to add the start:prod option to your package.json:
"scripts": {
"start": "react-scripts start",
"start:prod": "pushstate-server build",
"build": "react-scripts build",
"test": "react-scripts test --env=jsdom",
"eject": "react-scripts eject"
},
You can run the tests on your CI service with:
docker run <image> npm test
There's nothing stopping you from running this docker container locally as well to make sure things work as expected.

I recently made a small project called hello-docker-react who just does what the op is looking for.
It's made with docker-compose, create-react-app, yarn, a node image, and a small entrypoint script.
Live reload work flawlessly and I haven't found any problems yet.
https://github.com/lopezator/hello-docker-react

here is good gide for this
https://mherman.org/blog/dockerizing-a-react-app/
for development
# base image
FROM node:9.6.1
# set working directory
RUN mkdir /usr/src/app
WORKDIR /usr/src/app
# add `/usr/src/app/node_modules/.bin` to $PATH
ENV PATH /usr/src/app/node_modules/.bin:$PATH
# install and cache app dependencies
COPY package.json /usr/src/app/package.json
RUN npm install --silent
RUN npm install react-scripts#1.1.1 -g --silent
# start app
CMD ["npm", "start"]
for production
# build environment
FROM node:9.6.1 as builder
RUN mkdir /usr/src/app
WORKDIR /usr/src/app
ENV PATH /usr/src/app/node_modules/.bin:$PATH
COPY package.json /usr/src/app/package.json
RUN npm install --silent
RUN npm install react-scripts#1.1.1 -g --silent
COPY . /usr/src/app
RUN npm run build
# production environment
FROM nginx:1.13.9-alpine
COPY --from=builder /usr/src/app/build /usr/share/nginx/html
EXPOSE 80
CMD ["nginx", "-g", "daemon off;"]

Not exactly a direct improvement of the author's code, but I was able to get a development environment working with very little code - and no direct dependency to node on my machine - like this:
docker-compose.yml
services:
node:
image: node:16
user: "node"
command: "npm start"
working_dir: /app
volumes:
- ./:/app
ports:
- 3000:3000
This way, you avoid creating docker images from a Dockerfile.
Usage is generally like this:
install dependencies before running: docker compose run node npm install
run development environment: docker compose up
install new dependencies: docker compose run node npm install [package name]
clean up docker instances created with compose run: docker compose rm

While using docker in development with create-react-app, i discovered that it is possible to override the webpackDevServer configuration by adding CHOKIDAR_USEPOLLING=1to your .env file. This will make the file watching work again. It even refreshes the browser page on the host! The only thing that i discovered is that it doesn't open up a webpage automatically.
I can also advise to add tty: true to your service to have your original console output back into your terminal. To remove the container name prefixes in the logs, you can run something like this after running docker-compose up -d:
docker-compose logs -f --tail=100 client | cut -f2 -d \"|\""

Running with CRA 4.0 and many dependencies
.dockerignore
.git
.gitignore
node_modules
build
Dockerfile.dev
FROM node:alpine
WORKDIR /app
COPY package.json /app
RUN yarn install
COPY . .
CMD ["yarn", "start"]
docker-compose.dev.yml
version: "3.8"
services:
print:
stdin_open: true
build:
context: .
dockerfile: Dockerfile.dev
ports:
- "3000:3000"
volumes:
- ".:/app"
- "/app/node_modules"
Dockerfile.prod
FROM node:alpine as build
WORKDIR /app
COPY package.json /app
RUN yarn install
COPY . /app
RUN yarn run build
FROM nginx:stable-alpine
COPY ./nginx/nginx.conf /etc/nginx/conf.d/default.conf
COPY --from=build /app/build /usr/share/nginx/html
docker-compose.prod.yml
version: "3.8"
services:
print:
stdin_open: true
build:
context: .
dockerfile: Dockerfile.prod
ports:
- "80:80"
nginx.conf
server {
listen 80;
server_name frontend;
location / {
root /usr/share/nginx/html;
index index.html;
try_files $uri /index.html;
}
}
To run
docker-compose.exe -f .\docker-compose.yml up --build
or
docker-compose.exe -f .\docker-compose.dev.yml up --build

Here is a simple (pure docker) solution without local installation of runtime (e.g. node):
cd /tmp
docker run -it --rm -v "$PWD":/app -w /app node yarn create react-app my-app
sudo chown -R $USER:root my-app/
cd my-app
nano docker-compose.yml # see docker-compose.yml below
docker compose up -d
docker-compose.yml:
services:
node:
image: node:16-alpine
environment:
- CHOKIDAR_USEPOLLING=true
- FAST_REFRESH=true
working_dir: /app
ports:
- '3000:3000'
command: "yarn start"
volumes:
- './:/app'
open localhost:3000 in your browser. Hot reload should work out of the box.

Related

Docker Compose compiles successfully, and from within the container its returning the webpage successfully using curl but not from host

Im building out a docker dev environment and moving a bunch of existing apps into docker containers.
I have a an old react app which compiles using webpack, it seems no matter what I do in the docker compose file that it still remains unreachable from my host machine (MacOS)
running this works successfully, but not from the host
$ docker exec -it <container id> /bin/sh
# curl localhost:8080
<!DOCTYPE html>
<html>
...
</html>
Dockerfile
# syntax=docker/dockerfile:1
FROM node:16.16.0-buster as base
WORKDIR /app
ENV PATH /app/node_modules/.bin:$PATH
COPY package.json ./
COPY yarn.lock ./
COPY webpack.common.js ./
## import prod, because after install the build script tests if it can compile successfully
COPY webpack.prod.js ./
COPY . ./
RUN yarn install
EXPOSE 8080
FROM base as dev
CMD ["yarn", "run", "dev"]
Option 1) Doesnt Work
version: '3.8'
services:
backbone-ui:
network_mode: host
build:
context: .
container_name: backbone-ui
ports:
- target: 8080
host_ip: 127.0.0.1
published: 8080
protocol: http
mode: host
volumes:
- ./:/app
command: yarn run dev
Option 2) Doesnt Work
version: '3.8'
services:
backbone-ui:
network_mode: host
build:
context: .
container_name: backbone-ui
ports:
- 8080:8080
volumes:
- ./:/app
command: yarn run dev
Option 3) Doesnt Work
version: '3.8'
services:
backbone-ui:
build:
context: .
container_name: backbone-ui
ports:
- 8080:8080
volumes:
- ./:/app
command: yarn run dev
Failed Attempt
Dockerfile (paired with Option 2 compose.yml)
# syntax=docker/dockerfile:1
FROM node:16.16.0-buster as base
WORKDIR /app
ENV PATH /app/node_modules/.bin:$PATH
COPY package.json ./
COPY yarn.lock ./
COPY webpack.common.js ./
COPY webpack.prod.js ./
COPY . ./
RUN yarn install
FROM base as dev
EXPOSE 8080
CMD ["yarn", "run", "dev"]
Failed Attempt
updated package.json script
"dev": "webpack-dev-server --open --config webpack.dev.js --hot --host 0.0.0.0"
confirmed after rebuilding that the app is listening on 0.0.0.0:8080, but failed to respond on the host
$ docker exec -it <container id> /bin/sh
# curl 0.0.0.0:8080
<!DOCTYPE html>
<html>
...
</html>

React app refuses connection when dockerised

version: "1.0.0"
services:
########################################################################################################
############################################# VALIDLY #################################################
########################################################################################################
validly-studio:
build:
context: ./studio
dockerfile: Dockerfile
volumes:
- type: bind
source: ./studio
target: /app
- /app/node_modules
restart: unless-stopped
ports:
- 3000:3000
networks:
- validly
networks:
validly:
above is my docker-compose.yml file
FROM node:16.14-alpine
# set working directory
WORKDIR /app
# install app dependencies
COPY package.json ./
COPY package-lock.json ./
RUN npm install
# add app
COPY . ./
# start app
CMD ["npm", "start"]
this is my Dockerfile.
docker builds the react app and it prompts me to goto localhost:3000 where the app is running. But when I goto localhost:3000. I shows connection refused.
In your Dockerfile you copy in folder App everything to app/
And in your docker-compose you map /app to /app/node_modules.
This will not work.
Choose 1 of the two, and my instinct (and many error in the past) tell me that you should build everything in Dockerfile, including copying node_modules, and don't touch it in docker-compose.
Dockerfile
Template Dockerfile for React: (this one with NextJS environment, which makes it only more complex)
# Dependencies Container
FROM node:lts-alpine3.12 AS deps
RUN apk add --no-cache libc6-compat
WORKDIR /app
# Here we create node_modules
COPY package.json ./
COPY package-lock.json ./
RUN npm install -g npm#7.24.0 --no-update-notifier
RUN npm --version
RUN npm ci --no-update-notifier
# Rebuild the source code only when needed
FROM node:lts-alpine3.12 AS builder
WORKDIR /app
COPY . .
# Here we copy node_modules from previous intermediate container
COPY --from=deps /app/node_modules ./node_modules
RUN npm install -g npm#7.24.0 --no-update-notifier
RUN npm --version
RUN node -v
RUN npm run build --no-update-notifier
# Production Image
FROM node:16-bullseye AS runner
WORKDIR /app
ENV NODE_ENV production
# Here we only copy. No building needed, keeps the image small.
COPY --from=builder /app/.next ./.next
COPY --from=builder /app/node_modules ./node_modules
COPY --from=builder /app/package.json ./package.json
COPY --from=builder /app/package-lock.json ./package-lock.json
RUN addgroup -gid 1001 nodejs
RUN adduser -uid 1002 nextjs
RUN adduser nextjs nodejs
RUN chown -R nextjs:nodejs /app/.next
USER nextjs
docker-compose.yml
Here a template docker-compose.yml using the above Dockerfile
version: "3.9"
services:
webshop:
build:
context: ./build
dockerfile: Dockerfile_webshop
image: mywebshop
restart: "no"
container_name: MyWebshop
command: ["npm", "start"]
As you see, no volumes needed.

npm run tests in docker with travis CI

I am deploying a react app to Heroku via TravisCI. The fact that I'm using Heroku doesn't really affect what I'm about to ask, I'm pretty sure, it's just there for context. Travis successfully deploys the app until I add a testing step (the script section) in .travis.yml:
language: generic
sudo: required
services:
- docker
before_install:
- docker build -t myapp:prod -f Dockerfile.prod .
script:
- docker run -e CI=true myapp:prod npm run test
after_success:
- docker build -t myapp:prod -f Dockerfile.prod .
- echo "$DOCKER_PASSWORD" | docker login -u "$DOCKER_ID" --password-stdin
- docker push myapp:prod
deploy:
provider: heroku
app: myapp
skip_cleanup: true
api_key:
secure: <my_key>
However, my Dockerfile.prod is a multi-stage node + nginx where the nginx stage doesn't keep any node or npm stuff:
# build environment
FROM node:13.12.0-alpine as builder
# set working directory
WORKDIR /app
# add `/app/node_modules/.bin` to $PATH
ENV PATH /app/node_modules/.bin:$PATH
# install app dependencies
COPY package.json ./
COPY package-lock.json ./
# some CI stuff I guess
RUN npm ci
RUN npm install react-scripts#3.4.1 -g --silent
COPY . ./
RUN npm run build
# production environment
FROM nginx:stable-alpine
COPY nginx.conf /etc/nginx/conf.d/default.conf
# If using React Router
COPY --from=builder /app/build /usr/share/nginx/html
# For Heroku
CMD sed -i -e 's/$PORT/'"$PORT"'/g' /etc/nginx/conf.d/default.conf && nginx -g 'daemon off;'
Therefore, it is my understanding that .travis.yml tries to run that npm run test command inside my nginx container and can't execute npm commands (no node installed, right?). So guided by SO answers such as this one I started adding commands into that nginx stage such as
COPY package.json ./
COPY package-lock.json ./
RUN apk add --update npm
but I realized I might be approaching this the wrong way. Should I perhaps be adding npm through Travis? That is, should I include in .travis.yml in the scripts section something like docker run -e CI=true myapp:prod apk add --update npm and whatever else is necessary? This would result in a smaller nginx image no? However, would I run into problems with package.json from the node stage in Dockerfile.prod or anything like that?
In summary, to use TravisCI to test a dockerized react app served with nginx, at what point should I install npm into my image? Does it happen as part of script in .travis.yml or does it happen in Dockerfile.prod? If it is recommened to npm run tests inside Dockerfile.prod, would I do that in the first stage (node) or the second (nginx)?
Thanks
EDIT: Not sure if this can be considered solved, but a user on Reddit recommended to simply RUN npm run test right before the RUN npm run build.

Dockerized react app differences between running the image with and without docker-compose

// Dockerfile
# pull base image
FROM node:10.15.1-alpine
# set working directory
WORKDIR /app
# add `/app/node_modules/.bin` to $PATH
ENV PATH /app/node_modules/.bin:$PATH
# install app dependencies
COPY ./package.json ./
COPY ./yarn.lock ./
RUN yarn install
RUN yarn global add react-scripts#3.4.1 // not sure if this is even necessary
# add app
COPY . ./
# start app
CMD ["yarn", "start"]
Command I'm running to build it:
docker build -t matrix-fe:dev .
Command to run it:
docker run -it --rm -v ${PWD}:/app -v /app/node_modules -p 3000:3000 -e CHOKIDAR_USEPOLLING=true matrix-fe:dev
And then this is my composer yml:
version: '3.7'
services:
matrix-fe:
container_name: matrix-fe
build:
context: ./a-fe
dockerfile: Dockerfile
volumes:
- './a-fe:/app'
- '/app/node_modules'
ports:
- 3000:3000
environment:
- CHOKIDAR_USEPOLLING=true
Then to build and run it:
docker-compose up --build
Error I'm getting:
matrix-fe | yarn run v1.13.0
matrix-fe | $ react-scripts start
matrix-fe | It looks like you're trying to use TypeScript but do not have typescript installed.
matrix-fe | Please install typescript by running yarn add typescript.
matrix-fe | If you are not trying to use TypeScript, please remove the tsconfig.json file from your package root (and any TypeScript files).
matrix-fe |
Why is this happening? I can obviously try to also install typescript, but it is a dependency in package.json and should be installed, I also added node_modules to path. How is running the image with and without docker-compose different? Compose does create a different image called matrix_matrix-fe, but then Dockerfile hasn't changed. docker-compose.yml is in a top level folder, structure looks like this:
/matrix
/a-fe
./package.json
./Dockerfile
...
/a-be
./docker-compose.yml
Help me understand what's different, volumes are same, ENV variables are same, not seeing anything off.
Edit: forgot to mention that running the image without docker-compose doesn't output errors, it's working properly.

Docker container works locally but not when uploaded to elasticbeanstalk

My docker container works locally, I'm trying to deploy it on elastic beanstalk using travis.
My travis build is successful. The docker container has been tested locally and it works. On AWS Elastic Beanstalk I get a "Not a file/Directory error" for my build directory.
Dockerfile
FROM node:alpine as builder
WORKDIR '/app'
COPY package.json .
RUN npm install
COPY . .
CMD ["npm", "run", "build"]
#Run Phase
FROM nginx
EXPOSE 80
COPY --from=builder /app/build /usr/share/nginx/html
Dockerfile.dev
FROM node:alpine
WORKDIR '/app'
COPY package.json .
RUN npm install
COPY . .
CMD ["npm", "run", "start"]
travis.yml
sudo: required
services:
- docker
before_install:
- docker build -t *******/docker -f Dockerfile.dev .
script:
- docker run -e CI=true *******/docker npm run test -- --coverage
deploy:
provider: elasticbeanstalk
region: "ap-south-1"
app: "docker"
env: "Docker-env-2"
bucket_name: "***********************"
bucket_path: "docker"
on:
branch: master
access_key_id: $AWS_ACCESS_KEY
secret_access_key: $AWS_SECRET_KEY
Following are the logs -
Travis output
Elastic Beanstalk output
Any help would be appreciated, thanks!
To run it locally, I run the following commands-
1) docker build -t *******/docker .
2) docker run -it <port>:80 <container_id>
It works as expected and I can reach the server on localhost:.
I've put the same commands on the travis.yml file as well.
There are two dockerfiles because I would only be needing the "build" directory in the production container and I can ignore the rest of the directories to save space.
I realized that the build directory was listed in the .gitignore file, thereby preventing travis-ci from accessing it as it isn't in the repo.
Once I removed it and re-deployed it, worked perfectly.

Resources