зеркало из https://github.com/microsoft/docker.git
New Terminology section and illustrations.
This commit is contained in:
Родитель
0ae778c881
Коммит
a780b7c6b5
Двоичный файл не отображается.
После Ширина: | Высота: | Размер: 104 KiB |
Двоичный файл не отображается.
После Ширина: | Высота: | Размер: 77 KiB |
Двоичный файл не отображается.
После Ширина: | Высота: | Размер: 92 KiB |
Двоичный файл не отображается.
После Ширина: | Высота: | Размер: 76 KiB |
Разница между файлами не показана из-за своего большого размера
Загрузить разницу
После Ширина: | Высота: | Размер: 190 KiB |
|
@ -0,0 +1,18 @@
|
|||
:title: Terms
|
||||
:description: Definitions of terms used in Docker documentation
|
||||
:keywords: concepts, documentation, docker, containers
|
||||
|
||||
|
||||
|
||||
Terms
|
||||
=====
|
||||
|
||||
Definitions of terms used in Docker documentation.
|
||||
|
||||
Contents:
|
||||
|
||||
.. toctree::
|
||||
:maxdepth: 1
|
||||
|
||||
terms
|
||||
|
|
@ -0,0 +1,97 @@
|
|||
:title: Image & Container
|
||||
:description: Definitions of an image and container
|
||||
:keywords: containers, lxc, concepts, explanation, image, container
|
||||
|
||||
File Systems
|
||||
============
|
||||
|
||||
.. image:: images/docker-filesystems-generic.png
|
||||
|
||||
In order for a Linux system to run, it typically needs two `file
|
||||
systems <http://en.wikipedia.org/wiki/Filesystem>`_:
|
||||
|
||||
1. boot file system (bootfs)
|
||||
2. root file system (rootfs)
|
||||
|
||||
The **boot file system** contains the bootloader and the kernel. The
|
||||
user never makes any changes to the boot file system. In fact, soon
|
||||
after the boot process is complete, the entire kernel is in memory,
|
||||
and the boot file system is unmounted to free up the RAM associated
|
||||
with the initrd disk image.
|
||||
|
||||
The **root file system** includes the typical directory structure we
|
||||
associate with Unix-like operating systems: ``/dev, /proc, /bin, /etc,
|
||||
/lib, /usr,`` and ``/tmp`` plus all the configuration files, binaries
|
||||
and libraries required to run user applications (like bash, ls, and so
|
||||
forth).
|
||||
|
||||
While there can be important kernal differences between differnt Linux
|
||||
distributions, the contents and organization of the root file system
|
||||
are usually what make your software packages dependant on one
|
||||
distribution versus another. Docker can help solve this problem by
|
||||
running multiple distributions at the same time.
|
||||
|
||||
.. image:: images/docker-filesystems-busyboxrw.png
|
||||
|
||||
Layers and Union Mounts
|
||||
=======================
|
||||
|
||||
In a traditional Linux boot, the kernel first mounts the root file
|
||||
system as read-only, checks its integrity, and then switches the whole
|
||||
rootfs volume to read-write mode. Docker does something similar,
|
||||
*except* that instead of changing the file system to read-write mode,
|
||||
it takes advantage of a `union mount
|
||||
<http://en.wikipedia.org/wiki/Union_mount>`_ to add a read-write file
|
||||
system *over* the read-only file system. In fact there may be multiple
|
||||
read-only file systems stacked on top of each other.
|
||||
|
||||
.. image:: images/docker-filesystems-debianrw.png
|
||||
|
||||
At first, the top layer has nothing in it, but any time a process
|
||||
creates a file, this happens in the top layer. And if something needs
|
||||
to update an existing file in a lower layer, then the file gets copied
|
||||
to the upper layer and changes go into the copy. The version of the
|
||||
file on the lower layer cannot be seen by the applications anymore,
|
||||
but it is there, unchanged.
|
||||
|
||||
We call the union of the read-write layer and all the read-only layers
|
||||
a **union file system**.
|
||||
|
||||
Image
|
||||
=====
|
||||
|
||||
In Docker terminology, the read-only layer is called the
|
||||
**image**. An image never changes. Because Docker uses a union file
|
||||
system, the applications think the whole file system is mounted
|
||||
read-write, because any file can be changed. But all the changes go to
|
||||
the top-most layer, and underneath, the image is unchanged. Since they
|
||||
don't change, images do not have state.
|
||||
|
||||
Each image may depend on one more image which forms the layer beneath
|
||||
it. We sometimes say that the lower image is the **parent** of the
|
||||
upper image.
|
||||
|
||||
Base Image
|
||||
==========
|
||||
|
||||
An image that has no parent is a **base image**.
|
||||
|
||||
Container
|
||||
=========
|
||||
|
||||
Once you start a Docker container from an image, Docker fetches the
|
||||
image and its parent, and repeats the process until it reaches the
|
||||
base image. Then the union file system adds a read-write layer on
|
||||
top. That read-write layer, plus the information about its parent and
|
||||
some additional information like its unique id, is called a
|
||||
**container**.
|
||||
|
||||
Containers can change, and so they have state. A container may be
|
||||
running or exited. In either case, the state of the file system and
|
||||
its exit value is preserved. You can start, stop, and restart a
|
||||
container. The processes restart from scratch (their memory state is
|
||||
**not** preserved in a container), but the file system is just as it
|
||||
was when the container was stopped.
|
||||
|
||||
You can promote a container to an image with ``docker commit``. Once a
|
||||
container is an image, you can use it as a parent for new containers.
|
|
@ -18,5 +18,6 @@ This documentation has the following resources:
|
|||
contributing/index
|
||||
api/index
|
||||
faq
|
||||
terms/index
|
||||
|
||||
.. image:: concepts/images/lego_docker.jpg
|
||||
|
|
Загрузка…
Ссылка в новой задаче