Here is my setup:

• 4 Services: Worker, Database (Postgis), Rest-API and a Mapserver
• 1 Volume for the Database

docker volume inspect pgdata
[
   {
       "CreatedAt": "0001-01-01T00:00:00Z",
       "Driver": "local-persist",
       "Labels": {},
       "Mountpoint": "/mnt/pgdata",
       "Name": "pgdata",
       "Options": {
           "mountpoint": "/mnt/pgdata"
       },
       "Scope": "local"
   }
]

2 disks/partitions:

 df -h
Filesystem      Size  Used Avail Use% Mounted on
udev             63G     0   63G   0% /dev
tmpfs            13G  856K   13G   1% /run
/dev/sda1       939G   10G  889G   2% /
tmpfs            63G     0   63G   0% /dev/shm
tmpfs           5.0M     0  5.0M   0% /run/lock
tmpfs            63G     0   63G   0% /sys/fs/cgroup
s3fs             16E     0   16E   0% /codede
tmpfs            13G     0   13G   0% /run/user/1000
/dev/sdb1        46T   24K   44T   1% /mnt/pgdata

As you can see the partition mounted at /mnt/pgdata has 44TB storage and I want all the data stored in the Database to be stored there as it will get huge really fast.

But when I start to add data to the database, the 1TB disk will get filled subsequently and my system will fail/stop.

  database:
    image: DB_IMAGE
    volumes:
      - pgdata:/var/lib/postgresql/data
    command: postgres -c 'max_connections=500'
    ports:
      - "6543:5432"
    secrets:
      - postgres-user
      - postgres-password
    environment:
      POSTGRES_USER_FILE: /run/configs/postgres-user
      POSTGRES_PASSWORD_FILE: /run/secrets/postgres-password
      POSTGRES_DB: ts
    deploy:
      placement:
        constraints: [node.role == manager]

The workers are running on the nodes, mapserver and Rest-API are expected to have really small influence on the harddisk. So the only service running affecting the storage should be the database.

Why would it still fill the smaller diskspace instead of only the big one?

CodePudding user response：

Did you describe volume inside docker-compose.yml?

It is should be described even if already created externally:

version: "3.9"

# service description
services:
  frontend:
    image: node:lts
    volumes:
      - myapp:/home/node/app

# volume description
volumes:
  myapp:
    external: true

Or

volumes:
  pgdata:
    external: true

In your case

UPDATE ANSWER ACCORDING INFORMATION IN COMMENTS:

So apparently inside my /var/lib/docker-directory there is a subfolder with 125GB of disk space usage

/var/lib/docker/volumes is a default path for docker volumes.

Looks like your container using different pgdata volume. Maybe you not correctly described it (see my answer above). Here is hard to guess what happened according provided information.

If volume is really correctly described and still don't working I may recommend two workarounds:

1. Use direct mount without volumes:

   volumes:
     - /mnt/pgdata:/var/lib/postgresql/data
   

2. Mount second drive to /var/lib/docker to store all docker data at big disk by default:

   # stop docker
   service docker stop
   # unmount current folder
   umount /mnt/pgdata
   # rename old docker folder
   mv /var/lib/docker /var/lib/docker_old
   # create new folder
   mkdir /var/lib/docker
   # mount
   mount /dev/sdb1 /var/lib/docker
   # copy docker files to new position
   cp -pr /var/lib/docker_old/* /var/lib/docker/
   # check for hidden files starting from `.` if exists and copy it too
   ls -la /var/lib/docker_old/
   # remove old folder
   rm -fr /var/lib/docker_old
   # mount /dev/sdb1 to /var/lib/docker as permanent storage using /etc/fstab file
   # start docoker and enjoy
   service docker start