Docker-数据量和数据容器之间的区别

Question

i do not understand what difference there is between data volume and data containers... For example this two docker-compose confguration have same result but the first use a data container

whit data container

datacontainer:
 image: httpd:2.4-alpine
 volumes:
  - ../src:/usr/local/apache2/htdocs

apache:
 image: httpd:2.4-alpine
 ports:
  - 80:80
 volumes_from:
  - datacontainer

without data container

apache:
 image: httpd:2.4-alpine
 ports:
  - 80:80
 volumes:
  - ../src:/usr/local/apache2/htdocs

what are the advantages and disadvantages of using a data container?

Answer 1

Neither one of those examples are really data volumes or data containers. They are both bind mounts to the host, one more indirectly than the other. I tend to refer to those as host volumes.

A data container is deprecated to the named volumes. They would look like:

datacontainer:
 image: httpd:2.4-alpine
 command: tail -f /dev/null
 volumes:
  - /data

apache:
 image: httpd:2.4-alpine
 ports:
  - 80:80
 volumes_from:
  - datacontainer

Then you could recreate the apache container, or upgrade it, without upgrading the datacontainer, and still have your data available. There are multiple drawbacks to this, and if implemented with the volume defined in the image I posted a blog about just how much I dislike that . But the biggest issue is that you're managing your data as a container, so any container cleanup routines will equally cleanup your data, not very good if your opposed to data loss or container sprawl.

A named volume is much easier and has all the same features of a data container, but separates the data management from the container management, allowing containers to be purged without fear, and data to be backed up from one place. They look like:

apache:
 image: httpd:2.4-alpine
 ports:
  - 80:80
 volumes:
  - data:/data

Note in version 2 of compose, you'd be more specific with the volume definition.

Answer 2

The data container pattern was designed around data persistence, though it is mostly obsoleted by named volumes in Docker 1.9. The main advantage of both volumes and the data container pattern is that bind mounting on a host is host dependent, meaning you couldn't use that in a docker file. Volumes allow you the flexibility to define volumes when you build your images.

In addition, Data volumes are designed to persist data, independent of the container's life cycle. Docker therefore never automatically deletes volumes when you remove a container, nor will it “garbage collect” volumes that are no longer referenced by a container.