also, i’m running into exactly this 😅… https://github.com/calcom/docker/issues/121 followed the steps to build and did

DOCKER_BUILDKIT=0 docker compose build calcom

(all on my dev machine)… one question i have is… does

docker

somehow need a connection to the database while building the image? /cc @Colin - Krumware

You should be able to just use those variables as runtime

Yes, a database is required during build due to Prisma and the schema generation, also the migration scripting is currently part of the build. It does not have to be the same database as the one you use at runtime

hmm. i really can’t seem to get the custom docker image building instructions to work. everything works up until the command:

docker compose up -d database

then, with my

.env

configured to:

POSTGRES_USER=postgres
POSTGRES_PASSWORD=postgres
POSTGRES_DB=calendso
DATABASE_HOST=database:5432
DATABASE_URL=postgresql://${POSTGRES_USER}:${POSTGRES_PASSWORD}@${DATABASE_HOST}/${POSTGRES_DB}

i do:

DOCKER_BUILDKIT=0 docker compose build calcom

and end up with the error:

@calcom/prisma:build: Error: P1001: Can't reach database server at `database`:`5432`
@calcom/prisma:build: 
@calcom/prisma:build: Please make sure your database server is running at `database`:`5432`.
@calcom/prisma:build: error Command failed with exit code 1.
@calcom/prisma:build: info Visit <https://yarnpkg.com/en/docs/cli/run> for documentation about this command.
@calcom/prisma:build: error Command failed with exit code 1.
@calcom/prisma:build: info Visit <https://yarnpkg.com/en/docs/cli/run> for documentation about this command.

everything else about the project is stock. any thoughts here?

OK, after spending a few hours on it, i think i figured out a fix: 1. Currently, internal creation of

DATABASE_URL

value from the env for the prisma build step seems to be broken. Notice above, where the error message says ‘Error: P1001: Can’t reach database server at `database`:`5432`’. (The additional backtick marks seems suspicious.) What DOES work is passing in this value via the command line, like:

DOCKER_BUILDKIT=0 docker compose build --build-arg DATABASE_URL=<postgresql://postgres:postgres@host:5432/database> calcom

. (I tested setting the URL component values in the env to the same values as passed in via the command line like above: The env method failed and the command line method worked.) 2. The LAN IP address of the

database

container needs to be used at the command line, not the generic

database:5432

addressing. That is, before calling the above command, the user needs to

docker-compose up database --detach

,

docker exec -it database bash

, inside the

database

container get the LAN IP address via

hostname -I

, and then use that value as the appropriate URL host address within the command above. Example:

DOCKER_BUILDKIT=0 docker compose build --build-arg DATABASE_URL=<postgresql://postgres:postgres@172.24.0.2:5432/database> calcom

health checks look like they're configured here https://github.com/calcom/cal.com/blob/main/packages/prisma/docker-compose.yml

one note from someone new to Docker and an M1 Mac user… this change is required when building a custom image: https://stackoverflow.com/a/73400257

maybe it’s common knowledge to experts, but it’s definitely not to new folks. maybe worth calling that out in the README.md?

Hi guys, I am trying to deploy calcom in production, but when we execute the dockerfile using the yarn start, Dockerfile returns a 0 task total result.

zamas_next  | + scripts/replace-placeholder.sh  
zamas_next  | Nothing to replace, the value is already set to .
zamas_next  | + yarn build
zamas_next  | yarn run v1.22.19
zamas_next  | $ turbo run build --filter=@calcom/web...
zamas_next  |  WARNING  cannot find a .git folder. Falling back to manual file hashing (which may be slower). If you are running this build in a pruned directory, you can ignore this message. Otherwise, please initialize a git repos
itory in the root of your monorepo
zamas_next  | • Packages in scope: 
zamas_next  | • Running build in 0 packages
zamas_next  | 
zamas_next  |  Tasks:    0 successful, 0 total
zamas_next  | Cached:    0 cached, 0 total
zamas_next  |   Time:    2.963s
zamas_next  | 
zamas_next  | Done in 5.17s.
zamas_next  | yarn run v1.22.19
zamas_next  | $ turbo run start --scope="@calcom/web"
zamas_next  | • Packages in scope:
zamas_next  | 
zamas_next  |  Tasks:    0 successful, 0 total
zamas_next  | Cached:    0 cached, 0 total
zamas_next  |   Time:    107ms

Has anyone experienced this issue? This is my dockerfile

FROM node:16 as builder

RUN mkdir -p /app

WORKDIR /app
ARG NEXTAUTH_SECRET=secret
ARG CALENDSO_ENCRYPTION_KEY=secret
ARG NEXTAUTH_URL='<https://scheduling.hrbotfactory.com>'
ARG MAX_OLD_SPACE_SIZE=4096

ENV NEXTAUTH_SECRET=${NEXTAUTH_SECRET} \
    CALENDSO_ENCRYPTION_KEY=${CALENDSO_ENCRYPTION_KEY}

COPY package.json yarn.lock turbo.json ./
COPY packages ./packages
COPY apps/web ./apps/web


RUN yarn global add turbo && \
    yarn config set network-timeout 1000000000 -g && \
    turbo prune --scope=@calcom/web --docker && \
    yarn --global-folder /tmp/yarn/ install

#RUN yarn turbo run build --filter=@calcom/web

FROM node:16 as runner

WORKDIR /app

ARG NEXT_PUBLIC_WEBAPP_URL=<http://localhost:3000>

#ENV NODE_ENV production

RUN apt-get update && \
    apt-get -y install netcat && \
    rm -rf /var/lib/apt/lists/* && \
    npm install --global prisma

COPY package.json yarn.lock turbo.json ./
COPY --from=builder /app/node_modules ./node_modules
COPY --from=builder /app/packages ./packages
COPY --from=builder /app/apps/web ./apps/web
COPY --from=builder /app/packages/prisma/schema.prisma ./prisma/schema.prisma
COPY scripts scripts

RUN scripts/replace-placeholder.sh <http://NEXT_PUBLIC_WEBAPP_URL_PLACEHOLDER> ${NEXT_PUBLIC_WEBAPP_URL}

#ENV NODE_ENV development

RUN chmod 777 -R /app/node_modules/.cache/*
RUN chmod 777 -R /app/node_modules/.cache/turbo

COPY .env ./
COPY .env.example ./
COPY .env.appStore.example ./

COPY --from=builder --chown=nextjs:nodejs /app/apps/web/.next ./apps/web


EXPOSE 3000

CMD ["/app/scripts/start.sh"]

start.sh file:

#!/bin/sh
set -x

# Replace the statically built BUILT_NEXT_PUBLIC_WEBAPP_URL with run-time NEXT_PUBLIC_WEBAPP_URL
# NOTE: if these values are the same, this will be skipped.
scripts/replace-placeholder.sh "$BUILT_NEXT_PUBLIC_WEBAPP_URL" "$NEXT_PUBLIC_WEBAPP_URL"
yarn build
yarn start

@Fernando Sánchez did you initialize the git submodule? Step 3 here: https://github.com/calcom/docker#advanced-users-build-and-run-calcom

@Fernando Sánchez are you trying to build your own image or just run the public image?

Hi @Josh that was the problem, now I can go further but then, it thros the following error. => ERROR [runner 5/10] COPY --from=builder /node_modules ./node_modules 0.0s => ERROR [runner 6/10] COPY --from=builder /packages ./packages 0.0s => ERROR [runner 7/10] COPY --from=builder /apps/web ./apps/web 0.0s => ERROR [runner 8/10] COPY --from=builder /packages/prisma/schema.prisma ./prisma/schema.prisma 0.0s ------ > [runner 5/10] COPY --from=builder /node_modules ./node_modules: ------ ------ > [runner 6/10] COPY --from=builder /packages ./packages: ------ ------ > [runner 7/10] COPY --from=builder /apps/web ./apps/web: ------ ------ > [runner 8/10] COPY --from=builder /packages/prisma/schema.prisma ./prisma/schema.prisma: ------ failed to solve: failed to compute cache key: "/packages/prisma/schema.prisma" not found: not found

Hi self-hosters, any documentations about the server requirements? Something like how are RAM, CPU… for 100 CCU, 1000 CCU… Thanks

We dont have strict minimums published yet. The requirements are very low for running the image. If you're self-hosting Postgres then also include their recommended amounts in your considerations. I reserved 500ccu and 400mb RAM on my end for the calcom container on its own, think I can still go smaller

i’m currently hosting on lightsail and found it necessary to have 2GB RAM. i think that left a few hundred MB of RAM free at idle.

if you use the improved geoip csv, that alone will eat 1GB of RAM, IIRC (a comment from the maintainers is floating around the comments of one of the last 2 or 3 releases on github https://github.com/calcom/cal.com/tags)

I'm trying to build on Vultr and have found that the Cloud Compute (1 vCPU, 2 GB, 2.00 TB, 55 GB) VPS errors with heap memory overflow while doing a "yarn build".

@David Billsbrough 2GB of memory, yeah? I had to bump up Codespaces default to 8GB today not to get OOM error

YMMV ¯\_(ツ)_/¯

Yea the build is beefy because of all the turborepo stuff, unfortunately

just curious for my insight/tracking, what are y'all working on that requires the local build versus just pulling the image?

@Colin - Krumware Funny you should say that, because I ended up being forced to do that and was lucky I found the docker repo.

Thanks for all your hard work on that

ok sweet

do you want to add the render yaml to the docker repo directly?

It needs to be another repo

It tries to clone the repo and build. It won’t accept images.