Sapphire - Imagine a framework•12mo ago

Docker volume issue

Hi, trying to run my bot inside of a docker container on my server (exposing the workdir to the host for changes). Yes, these files are stolen from the official repos. Current bind shows no files on the host and the attached error when trying to run the container. If I run the container without any mount points then it runs fine.

# ================ #
#    Base Stage    #
# ================ #

FROM node:20-bullseye-slim as base

WORKDIR /usr/src/app

# ENV YARN_DISABLE_GIT_HOOKS=1
ENV CI=true
ENV LOG_LEVEL=info
ENV FORCE_COLOR=true

RUN apt-get update && \
    apt-get upgrade -y --no-install-recommends && \
    apt-get install -y --no-install-recommends build-essential python3 dumb-init && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/* && \
    apt-get autoremove

COPY --chown=node:node yarn.lock .
COPY --chown=node:node package.json .
COPY --chown=node:node .yarnrc.yml .
COPY --chown=node:node .yarn/ .yarn/

ENTRYPOINT ["dumb-init", "--"]

# ================ #
#   Builder Stage  #
# ================ #

FROM base as builder

# ENV NODE_ENV="development"

# COPY --chown=node:node tsconfig.base.json .
COPY --chown=node:node tsconfig.json .
# COPY --chown=node:node tsup.config.ts .
# COPY --chown=node:node prisma/ prisma/
COPY --chown=node:node src/ src/

RUN yarn install --immutable
# RUN yarn run build

# ================ #
#   Runner Stage   #
# ================ #

# FROM base AS runner

ENV NODE_ENV="production"
ENV NODE_OPTIONS="--enable-source-maps"

# COPY --chown=node:node src/.env src/.env
# COPY --chown=node:node --from=builder /usr/src/app/dist dist

# RUN yarn workspaces focus --all --production

# Patch .prisma with the built files
# COPY --chown=node:node --from=builder /usr/src/app/node_modules/.prisma node_modules/.prisma

RUN chown node:node /usr/src/app/

USER node

CMD [ "yarn", "run", "watch:start" ]

# ================ #
#    Base Stage    #
# ================ #

FROM node:20-bullseye-slim as base

WORKDIR /usr/src/app

# ENV YARN_DISABLE_GIT_HOOKS=1
ENV CI=true
ENV LOG_LEVEL=info
ENV FORCE_COLOR=true

RUN apt-get update && \
    apt-get upgrade -y --no-install-recommends && \
    apt-get install -y --no-install-recommends build-essential python3 dumb-init && \
    apt-get clean && \
    rm -rf /var/lib/apt/lists/* && \
    apt-get autoremove

COPY --chown=node:node yarn.lock .
COPY --chown=node:node package.json .
COPY --chown=node:node .yarnrc.yml .
COPY --chown=node:node .yarn/ .yarn/

ENTRYPOINT ["dumb-init", "--"]

# ================ #
#   Builder Stage  #
# ================ #

FROM base as builder

# ENV NODE_ENV="development"

# COPY --chown=node:node tsconfig.base.json .
COPY --chown=node:node tsconfig.json .
# COPY --chown=node:node tsup.config.ts .
# COPY --chown=node:node prisma/ prisma/
COPY --chown=node:node src/ src/

RUN yarn install --immutable
# RUN yarn run build

# ================ #
#   Runner Stage   #
# ================ #

# FROM base AS runner

ENV NODE_ENV="production"
ENV NODE_OPTIONS="--enable-source-maps"

# COPY --chown=node:node src/.env src/.env
# COPY --chown=node:node --from=builder /usr/src/app/dist dist

# RUN yarn workspaces focus --all --production

# Patch .prisma with the built files
# COPY --chown=node:node --from=builder /usr/src/app/node_modules/.prisma node_modules/.prisma

RUN chown node:node /usr/src/app/

USER node

CMD [ "yarn", "run", "watch:start" ]

version: '3.9'

services:
  lyra:
    build:
      context: ../
    volumes:
      - .:/usr/src/app
    container_name: lyra
    logging:
      options:
        max-size: '1g'
        max-file: '3'
    networks:
      - infi
    restart: always
    tty: true
    image: healthcareusa/lyra:latest

networks:
  infi:

version: '3.9'

services:
  lyra:
    build:
      context: ../
    volumes:
      - .:/usr/src/app
    container_name: lyra
    logging:
      options:
        max-size: '1g'
        max-file: '3'
    networks:
      - infi
    restart: always
    tty: true
    image: healthcareusa/lyra:latest

networks:
  infi:

Solution:

No but you can make the GH action that also published the new image to run that command on the server

Jump to solution

13 Replies

Favna•12mo ago

Please change that network name lmfao. That's an inside joke for me and me alone. and why do you want to use a volume map like that? You're supposed to build a new docker image (in GH CI) then pull and restart on the server. You're not supposed to have the source files on the server as well, at that point you may as well just use pm2 instead of docker. 1. push code 2. trigger GH action 3. this publishes a new version of the docker image to the github container registry 4. run docker compose pull lyra && docker compule up -d --force-recreate --no-deps lyra to pull and restart the bot on the server 5. ??? 6. profit!

SongbirdOP•12mo ago

I don't want to have to check every 5 minutes (or whatever) for updates using a separate docker container on my server trollformation

there a way to listen to updates? there probably is I'll look in to it

Solution

Favna•12mo ago

No but you can make the GH action that also published the new image to run that command on the server

Favna•12mo ago

I created https://github.com/favware/ssh-remote-action for that purpose. Currently only use it for https://github.com/favware/graphql-pokemon but in the future more.

GitHub

GitHub - favware/ssh-remote-action: A GitHub action for executing a...

A GitHub action for executing a SSH command on a remote server - GitHub - favware/ssh-remote-action: A GitHub action for executing a SSH command on a remote server

Favna•12mo ago

also you dont need to check every 5 minutes. You check when you push new code. There is no reason to check otherwise. when you've pushed new code and you want it to be released to the masses you update and restart the bot. Either don't automated it at all, or automate on every commit as mentioned above

SongbirdOP•12mo ago

https://tenor.com/view/ohhh-duh-why-didnt-i-think-of-that-gif-21849807

Tenor

SongbirdOP•12mo ago

this is why I am in a server with smart people

Favna•12mo ago

just 1 thing to keep in mind with ssh-remote-action, it wont have access to aliases and env vars and the like on the server that you load through .bashrc / .zshrc etc. So for the command you configure keep that in mind. also best to use absolute paths if you need to cd anything so no ~/ stuff for graphql-pokemon I have a script called control.sh at /root/workspace/control.sh so I call /root/workspace/control.sh update pokedex-v8 (update = in the switchcase of the script, pokedex-v8 is the service in docker-compose) but if you have just the 1 service just embed it as the commadn, no need to call a shell script

SongbirdOP•12mo ago

thank you!! @Favna am i correct in assuming the example workflow yml is the only one needed for building image/deploying it? just confirming that "publish" means build/push docker image confused as to why it is "on release" when the workflow is the one triggering a release? https://github.com/favware/graphql-pokemon/actions/runs/7334424446/workflow is this the relevant workflow for what I am trying to do? and how the F*** do I push to GHCR?!?

Favna•12mo ago

name: Continuous Deployment

on:
  push:
    branches:
      - main

jobs:
  publish:
    name: Publish image to container registries
    runs-on: ubuntu-latest
    steps:
      - name: Checkout Project
        uses: actions/checkout@v4
      - name: Setup Docker Buildx
        uses: docker/[email protected]
      - name: Login to GitHub Container Registry
        uses: docker/[email protected]
        with:
          registry: ghcr.io
          username: ${{ github.repository_owner }}
          password: ${{ secrets.GITHUB_TOKEN }}
      - name: Build and push Docker image
        uses: docker/[email protected]
        with:
          push: true
          context: .
          tags: ghcr.io/favware/graphql-pokemon:latest

  updateOnServer:
    name: Update running container on server
    runs-on: ubuntu-latest
    needs: publish
    steps:
      - name: Update container on server
        uses: favware/ssh-remote-action@v1
        with:
          host: ${{ secrets.SSH_HOST }}
          port: ${{ secrets.SSH_PORT }}
          key: ${{ secrets.SSH_KEY }}
          passphrase: ${{ secrets.SSH_KEY_PASSPHRASE }}
          username: ${{ secrets.SSH_USERNAME }}
          command: ${{ secrets.SSH_COMMAND }}
          silent: true

name: Continuous Deployment

on:
  push:
    branches:
      - main

jobs:
  publish:
    name: Publish image to container registries
    runs-on: ubuntu-latest
    steps:
      - name: Checkout Project
        uses: actions/checkout@v4
      - name: Setup Docker Buildx
        uses: docker/[email protected]
      - name: Login to GitHub Container Registry
        uses: docker/[email protected]
        with:
          registry: ghcr.io
          username: ${{ github.repository_owner }}
          password: ${{ secrets.GITHUB_TOKEN }}
      - name: Build and push Docker image
        uses: docker/[email protected]
        with:
          push: true
          context: .
          tags: ghcr.io/favware/graphql-pokemon:latest

  updateOnServer:
    name: Update running container on server
    runs-on: ubuntu-latest
    needs: publish
    steps:
      - name: Update container on server
        uses: favware/ssh-remote-action@v1
        with:
          host: ${{ secrets.SSH_HOST }}
          port: ${{ secrets.SSH_PORT }}
          key: ${{ secrets.SSH_KEY }}
          passphrase: ${{ secrets.SSH_KEY_PASSPHRASE }}
          username: ${{ secrets.SSH_USERNAME }}
          command: ${{ secrets.SSH_COMMAND }}
          silent: true

@HealthCareUSA this is broken down to the essential

SongbirdOP•12mo ago

thank you!! what keytype does this support?

Favna•12mo ago

Any your server does. It just uses ssh in the background.

SongbirdOP•12mo ago

interesting for some reason it isn't letting me use it, yet I use the same key for SFTP figured it out unsupported format lol converted to openssh format

Gaming

Programming

Docker volume issue