image 1.7.1-fpm gives public_html error

[nupulu]

Tried to use explicitly image 1.7.1-fpm and I get this:
Please, configure your HTTP server to point to the /public_html directory (with fallback to /public_html/index.php.

I thought this problem was fixed in 1.7.1, but I still see it.

Is there something I need to do on my end to get it back on track?

Side note: image latest-fpm still is 1.7.0-fpm, I tought it should be 1.7.1-fpm. Just my 2 cents.

[pabzm]

The latest-fpm image now also brings Roundcubemail v1.7.1.

You're the first to report this, for others it works as expected, so I'm suspecting a local issue.

Can you please explain how you use the image? And let us know which image it is exactly (the ID, please)?

[nupulu]

docker ps gives: cd195d61e710 roundcube/roundcubemail:latest-fpm

docker-compose.yaml:

services:
  roundcubemail:
    image: roundcube/roundcubemail:latest-fpm
#    image: roundcube/roundcubemail:1.7.x-fpm
    container_name: roundcubemail
    # restart: unless-stopped
    depends_on:
      - roundcubedb
    links:
      - roundcubedb
    volumes:
      - ./www:/var/www/html
      - ./config:/var/roundcube/config
    environment:
      - ROUNDCUBEMAIL_DB_TYPE=pgsql
      - ROUNDCUBEMAIL_DB_HOST=roundcubedb # same as pgsql container name
      - ROUNDCUBEMAIL_DB_NAME=roundcube # same as pgsql POSTGRES_DB env name
      - ROUNDCUBEMAIL_DB_USER=roundcube # same as pgsql POSTGRES_USER env name
      - ROUNDCUBEMAIL_DB_PASSWORD=roundcube # same as pgsql POSTGRES_PASSWORD env name
      - ROUNDCUBEMAIL_SKIN=elastic
      - ROUNDCUBEMAIL_DEFAULT_HOST=ssl://mail.my.server.bla
      - ROUNDCUBEMAIL_DEFAULT_PORT=993
      - ROUNDCUBEMAIL_SMTP_SERVER=ssl://mail.my.server.bla
      - ROUNDCUBEMAIL_SMTP_PORT=465
      - ROUNDCUBEMAIL_PLUGINS=archive,zipdownload, managesieve
      - ROUNDCUBEMAIL_ASPELL_DICTS=en,de

  roundcubedb:
    image: postgres:17
    container_name: roundcubedb
    # restart: unless-stopped
    volumes:
      - ./db/postgres:/var/lib/postgresql/data
      - ./backup:/backup
    environment:
      - POSTGRES_DB=roundcube
      - POSTGRES_USER=roundcube
      - POSTGRES_PASSWORD=roundcube

  roundcubenginx:
    image: nginx:latest
    container_name: roundcubenginx
    # restart: unless-stopped
    ports:
      - '80:80'   # host:container
      # If you need SSL connection
      - '443:443' # host:container
    depends_on:
      - roundcubemail
    links:
      - roundcubemail
    volumes:
      - ./www:/var/www/html
      - ./nginx/templates:/etc/nginx/templates
      - /mnt/containerdata-swag/config/etc/letsencrypt:/etc/letsencrypt

    environment:
      - NGINX_HOST=mail.my.server.bla  # set your local domain or your live domain
      - NGINX_PHP_CGI=roundcubemail:9000  # same as roundcubemail container name

[pabzm]

docker ps shows the running container ID, not the image ID.

Please run this command and show the output: docker image inspect docker.io/roundcube/roundcubemail:latest-fpm --format="{{ .Id }}"

Did you try to freshly pull the image tag?

[nupulu]

fresh pull, then docker image inspect docker.io/roundcube/roundcubemail:latest-fpm --format="{{ .Id }}: sha256:c78890999adca46eb3c9b46ef796f35c203f0a1283b366f54358d4a7eedc4c57

[nmonserrato]

I'm facing the same issue with latest-apache, but all works fine with 1.6.16-apache. Tried also removing the entire www and db folder for a fresh installation.

I'm using podman and this is my quadlet config:
https://github.com/nmonserrato/homelab-quadlets-main/blob/main/roundcube.container

If I open a shell into the pod and check the only sites-enabled conf, I see the DocumentRoot pointing to the correct public_html folder, I have no idea how to debug this. Any ideas?

[nupulu]

The problem is still there. Can we get help with this?

docker image inspect docker.io/roundcube/roundcubemail:latest-fpm --format="{{ .Id }}"
sha256:6d0c748c1574430b9f0e43e73d41b4997179165b60f8a0432f779c96d0817aa7

[pabzm]

Both of you have /var/www/html mounted as volume. Does the setup work if you run it without that?

[nupulu]

Both of you have /var/www/html mounted as volume. Does the setup work if you run it without that?

I commented that mount out and it does not help. Same error message.