Error whe adding an image to an exercise about docker HOT 8 CLOSED

I've commited a fix in master, but want to test whether the autoincrement ID needs to be reset on the database since it was being set manually in the download script (also, running the script will not work till the website is updated)

from docker.

Great! Closing here, if you encounter the problem again, just re-open

from docker.

That's interesting (and obviously a bug). It's also interesting that it clashes with an id when it should add a new one. Does the docker imagine have already exercise images?

from docker.

The docker image hasn't images, but, after using docker compose, I generate the images using the command

docker-compose exec web python3 manage.py download-exercise-images

docker-compose exec web wger load-online-fixtures

Without any error message.

I even tried to exec docker-compose exec web python3 manage.py migrate, but it didn't solved the issue.

I used the docker-compose up -d, in a fresh install, using the following docker-compose.yml and prod.env below.
I just changed the port (from 80 to 58148), and volumes, but I think the problem is not related to that.

docker-compose.yml

version: '3.7'

services:
  web:
    image: wger/devel:latest
    volumes:
      - /home/username/wger/static:/home/wger/static
      - /home/username/wger/media:/home/wger/media
    ports:
      - 8000
    env_file:
      - ./prod.env
    depends_on:
      - db
      - cache

  nginx:
    build: ./nginx
    volumes:
      - /home/username/wger/static:/wger/static:ro
      - /home/username/wger/media:/wger/media:ro
    ports:
      - 58148:80
    depends_on:
      - web

  db:
    image: postgres:12.0-alpine
    volumes:
      - /home/username/wger/postgres-data:/var/lib/postgresql/data/
    expose:
      - 5432
    environment:
      - POSTGRES_USER=wger
      - POSTGRES_PASSWORD=wger
      - POSTGRES_DB=wger

  cache:
    restart: always
    image: redis:latest
    expose:
      - 6379

volumes:
  postgres-data:
  static:
  media:

prod.env

# Database
DJANGO_DB_ENGINE=django.db.backends.postgresql
DJANGO_DB_DATABASE=wger
DJANGO_DB_USER=wger
DJANGO_DB_PASSWORD=wger
DJANGO_DB_HOST=db
DJANGO_DB_PORT=5432

# Cache
DJANGO_CACHE_BACKEND=django_redis.cache.RedisCache
DJANGO_CACHE_LOCATION=redis://cache:6379/1
DJANGO_CACHE_TIMEOUT= 1296000
# 60*60*24*15, 15 Days
DJANGO_CACHE_CLIENT_CLASS=django_redis.client.DefaultClient

DJANGO_MEDIA_ROOT=/home/wger/media

# Change the secret key to a random string if you are running this instance publicly
SECRET_KEY=mysecretkey


# Application
ALLOW_REGISTRATION=True
ALLOW_GUEST_USERS=False

# Others
DJANGO_DEBUG=False
WGER_USE_GUNICORN=True
TIME_ZONE=Europe/Berlin
SITE_URL=http://localhost:8000

from docker.

Found the error, the exercise images were being identified by their ID, so if there were locally uploaded images they would clash when (or after) running download-exercise-images. I'll add a UUID as we do with the exercises so it's possible to properly reference them

from docker.

Ok, I will wait for the tests. Thank you! Really nice software.

from docker.

you might have to change the postgres sequence for the image IDs. For that, you will need to exec into the DB container and do something like this:

ALTER SEQUENCE exercises_exerciseimage_id_seq RESTART WITH 500;

I'm also not sure whether you need to to reload the exercise and exercise base fixtures, probably not. So, depending on how much data you have entered, it might be faster to drop the DB and pull a new version
(while I was looking into this, I refactored the image script, it should be way faster now)

from docker.

Thanks! I was just starting to add pictures, so I will drop the DP and pull the new version.

from docker.