Creating polls (demo poll) doesnt work with Docker

[blazekjan]

Hello,
I have a trouble running Rallly in Docker. I tried to spin up the containers on two different machines, however I am still being stuck. What did I do:

1. edited docker-compose to have db-data in docker-compose DIR (as everything gets pulled and it is automatically backed up)
   • ./db-data:/var/lib/postgresql/data
2. edited .env, and commented out (as stated in readme)
   DATABASE_URL=postgres://your-database/db
   set up 32 characters long SECRET_PASSWORD
   set up SMTP, where I used same credentials as for vaultwarden which works
3. hit docker compose up -d

What do I get:

[+] Running 3/0
 ⠿ Network rallly_default        Created                                                                                                                                                                                                                                                                                                                            0.0s
 ⠿ Container rallly-rallly_db-1  Created                                                                                                                                                                                                                                                                                                                            0.0s
 ⠿ Container rallly-rallly-1     Created                                                                                                                                                                                                                                                                                                                            0.0s
Attaching to rallly-rallly-1, rallly-rallly_db-1
rallly-rallly_db-1  |
rallly-rallly_db-1  | PostgreSQL Database directory appears to contain a database; Skipping initialization
rallly-rallly_db-1  |
rallly-rallly_db-1  | 2022-05-23 21:50:58.112 UTC [1] LOG:  starting PostgreSQL 14.2 (Debian 14.2-1.pgdg110+1) on x86_64-pc-linux-gnu, compiled by gcc (Debian 10.2.1-6) 10.2.1 20210110, 64-bit
rallly-rallly_db-1  | 2022-05-23 21:50:58.112 UTC [1] LOG:  listening on IPv4 address "0.0.0.0", port 5432
rallly-rallly_db-1  | 2022-05-23 21:50:58.112 UTC [1] LOG:  listening on IPv6 address "::", port 5432
rallly-rallly_db-1  | 2022-05-23 21:50:58.112 UTC [1] LOG:  listening on Unix socket "/var/run/postgresql/.s.PGSQL.5432"
rallly-rallly_db-1  | 2022-05-23 21:50:58.114 UTC [25] LOG:  database system was shut down at 2022-05-23 21:50:50 UTC
rallly-rallly_db-1  | 2022-05-23 21:50:58.120 UTC [1] LOG:  database system is ready to accept connections
rallly-rallly-1     | yarn run v1.22.19
rallly-rallly-1     | $ /usr/src/app/node_modules/.bin/prisma migrate deploy --schema prisma/schema.prisma
rallly-rallly-1     | Prisma schema loaded from prisma/schema.prisma
rallly-rallly-1     | Datasource "db": PostgreSQL database "db", schema "public" at "rallly_db:5432"
rallly-rallly-1     |
rallly-rallly-1     | 17 migrations found in prisma/migrations
rallly-rallly-1     |
rallly-rallly-1     |
rallly-rallly-1     | No pending migrations to apply.
rallly-rallly_db-1  | 2022-05-23 21:51:00.681 UTC [33] LOG:  could not receive data from client: Connection reset by peer
rallly-rallly-1     | Done in 2.25s.
rallly-rallly-1     | yarn run v1.22.19
rallly-rallly-1     | $ next start
rallly-rallly-1     | ready - started server on 0.0.0.0:3000, url: http://localhost:3000

docker-compose.yml

version: "3.3"
services:
  rallly_db:
    image: postgres:14.2
    restart: always
    volumes:
      - ./db-data/:/var/lib/postgresql/data
    environment:
      - POSTGRES_PASSWORD=postgres

  rallly:
    build:
      context: .
      args:
        - DATABASE_URL=postgres://postgres:postgres@rallly_db:5432/db?pgbouncer=true
        - NEXT_PUBLIC_BASE_URL=${NEXT_PUBLIC_BASE_URL}
    restart: always
    command: sh -c "yarn prisma migrate deploy --schema prisma/schema.prisma && yarn start"
    depends_on:
      - rallly_db
    ports:
      - 3000:3000
    environment:
      - DATABASE_URL=postgres://postgres:postgres@rallly_db:5432/db
    env_file:
      - .env

volumes:
  db-data:
    driver: local

.env

NEXT_PUBLIC_BASE_URL=http://localhost:3000
#DATABASE_URL=postgres://your-database/db
SECRET_PASSWORD=*redacted*
SUPPORT_EMAIL=*redacted*
SMTP_HOST=smtp.sendgrid.net
SMTP_PORT=465
SMTP_SECURE=true
SMTP_USER=apikey
SMTP_PWD=*redacted*

I can't create any polls and demo poll doesnt work either. Its stuck on Creating demo poll and I get: Uh oh! Something went wrong. The issue has been logged and we'll fix it as soon as possible. Please try again later.

Just checked recently closed issues and found #178 Any chance it has something to do with?

[Daedalus3]

It looks like you missed setting NEXT_PUBLIC_BASE_URL so the API can't be reached.

Also I did set

DATABASE_URL=postgres://postgres:postgres@rallly_db:5432/db

even though the README suggests otherwise. That way it works fine.

[blazekjan]

It looks like you missed setting NEXT_PUBLIC_BASE_URL so the API can't be reached.

Also I did set

DATABASE_URL=postgres://postgres:postgres@rallly_db:5432/db

even though the README suggests otherwise. That way it works fine.

No I didn't, NEXT_PUBLIC_BASE_URL is included in the .env file. Database connection is specified in the docker-compose file. Database shouldn't be a problem.

[Daedalus3]

All right -your .env file and log suggested otherwise (set to localhost).

[blazekjan]

All right -your .env file and log suggest otherwise (set to localhost).

Of course it is set to localhost, as the service should be reachable on multiple IPs.

[lukevella]

We spoke yesterday about this and you said you will try on a different machine. Were you able to do that?

[lukevella]

Also can you open the network tab on your browser to see what the failed request is returning when you try to create a demo?

[blazekjan]

We spoke yesterday about this and you said you will try on a different machine. Were you able to do that?

Yeah, I tried that even with different docker (and docker compose) versions. Still unable to create any polls, not even demo.

[lukevella]

Ok well that error has nothing to do with this project. Your requests are being blocked by something. Usually it's an ad-blocker. Feel free to report back when you figure out what that is. I'd start with googling ERR_BLOCKED_BY_CLIENT.

[blazekjan]

I am suspecting the problem somewhere else, as firefox popped up something more interesting. If it has something to do with referrer policy, I might try to pass the service through reverse proxy and try again. It is weird behaviour, but whatever.

Cross-Origin Request Blocked: The Same Origin Policy disallows reading the remote resource at http://localhost:3000/api/trpc/polls.demo.create?batch=1. (Reason: CORS request did not succeed). Status code: (null).

[Daedalus3]

The browser won't find the remote source at localhost (but this is my last shot getting you to edit it ;-) ). That was exactly the problem in #178.

[blazekjan]

The browser won't find the remote source at localhost (but this is my last shot getting you to edit it ;-) ). That was exactly the problem in #178.

That would mean that bug #178 is not fixed at all. Thank you, I will look into it in a bit.

[lukevella]

#178 is fixed. Make sure that the URL your are accessing the app on (the one you type in to your browser) is the same URL being set in NEXT_PUBLIC_BASE_URL.

[blazekjan]

Thank you for this awesome project @lukevella. Thank you also @Daedalus3 for pointing me the right way. I managed to make it work, but I am not happy about the way how I did it and how it works after all.

I would consider changing the way how the docker image builds. I think docker builds should not reference the IP address or domain (same as #170 with db). You cannot change it afterwards and with every change you need to rebuild the whole image. This is not very clear in the build instructions and one of the reasons I failed too many times.

I had different approach. I thought that NEXT_PUBLIC_BASE_URL=http://localhost:3000 means, it listens on all available interfaces (which is a common way in many configs), but this is wrong. You cant even reverse proxy the service and upstream port 3000 to different address to make it work (or maybe you can - but you must pass the correct headers??). Anyways I couldnt manage it to work with caddy. But if you build and run the image it still says in the logs ready - started server on 0.0.0.0:3000, url: http://localhost:3000. Thats confusing!

We pass almost everything through reverse proxies these days, that would be a good enhancement and one less variable needed with more flexible configuration.