docker-compose with nginx and nextcloud FPM stops accepting https connections

Question

I have a fairly standard setup, with nginx as a web frontend (with https certiticates and such) and a nextcloud FPM backend; the whole installation has very low traffic, since I am the only one using it.

At one point https stops working, every connection from outside results in either timeout or connection refused; in this situation it seem that nextcloud is working and nginx isn't. I thought it may have to do with some energy saving on my host, but it's not the case, since all the other containers on the same host work just fine; the funny thing is that there is a simple way to make it work again and it is to open a shell on the host and do:

cd $NEXTCLOUD_DIRECTORY

where NEXTCLOUD_DIRECTORY is the base directory for the application, where both docker-compose.yml file and data directories are (located at $HOME/docker/nextcloud-letsencrypt).

I just don't understand why this is happening and why that operation is a solution...

Here is my compose file:

version: '3'
  services:
    nginx:
      image: nginx:alpine
      ports:
        - "80:80"
        - "127.0.0.1:8443:443"
      volumes:
        - ./data/nginx:/etc/nginx/conf.d
        - ./data/certbot/conf:/etc/letsencrypt
        - ./data/certbot/www:/var/www/certbot
        - ./data/nextcloud/www:/var/www/html:ro
        - ./data/nextcloud/apps:/var/www/html/custom_apps:ro
      restart: unless-stopped
      command: "/bin/sh -c 'while :; do sleep 6h & wait $${!}; nginx -s reload; done & nginx -g \"daemon off;\"'"
    certbot:
      image: certbot/certbot
      volumes:
        - ./data/certbot/conf:/etc/letsencrypt
        - ./data/certbot/www:/var/www/certbot
      entrypoint: "/bin/sh -c 'trap exit TERM; while :; do certbot renew; sleep 12h & wait $${!}; done;'"
      restart: unless-stopped
    cloud-db:
      container_name: ${DB_CONTAINER_NAME}
      image: mariadb:${DB_IMAGE_TAG}
      restart: unless-stopped
      command: --transaction-isolation=READ-COMMITTED --binlog-format=ROW
      volumes:
        ...
      environment:
        MYSQL_DATABASE: ${MYSQL_DATABASE}
        MYSQL_ROOT_PASSWORD: ${MYSQL_ROOT_PASSWORD}
        MYSQL_USER: ${MYSQL_USER}
        MYSQL_PASSWORD: ${MYSQL_PASSWORD}
    app:
     image: nextcloud:21-fpm-alpine
     links:
       - cloud-db
     user: "1000:1004"
     volumes:
       - ./data/nextcloud/www:/var/www/html
       - ./data/nextcloud/apps:/var/www/html/custom_apps
       - ./data/nextcloud/config:/var/www/html/config
       - /mnt/usb/shared/nextcloud:/var/www/html/data
       - /mnt/usb/Expansion_2/serie:/mnt/serie:ro
       - /mnt/usb/archivio/archivio:/mnt/archivio:ro
     restart: unless-stopped

./data is in the root FS, inside user 1000's home directory.

/mnt/usb/shared/nextcloud is on an ext4 USB drive (ext4 defaults,nofail 0 0), R/W to the user

/mnt/usb/Expansion_2 is another ex4 USB drive (ext4 defaults,nofail 0 0) served through NC (they are registered as external storage in NC)

Did you do something strange like encrypt your home directory? — Michael Hampton, Jun 23 '21 at 10:55
no, nothing like that; I actually noticed that the containers are recreated when I do the "cd" operation. They result created a minute ago or so. I think there's something wrong in my compose configuration rather than either nginx or nextcloud — Riccardo Cossu, Jun 23 '21 at 12:57

score 0 · Accepted Answer · answered Jun 30 '21 at 08:52

It seems that the solution is to use restart:always instead of "unless-stopped". I don't really know why I need that since no one is stopping the containers, so they should run indefinitely, like many other containers I have on the same machine that use "unless-stopped" and stay up indefinitely.

docker-compose with nginx and nextcloud FPM stops accepting https connections

1 Answers1