Docker data persistence and container migration

The last section mentions that when the container is running, the data generated will not be in the write image, and re-starting the new container with this image will initialize the image and add a new read-write layer to hold the data. If you want to do data persistence, Docker provides data volume or data container volumes to solve the problem, and commits a new image to hold the resulting data. So, let's look at each of the different ways to use it.

First, the data volume

Data volume Features:

• You can bypass the UFs file system and provide access to one or more containers.

• is completely independent of the container's life cycle, so it does not delete the data volume it hangs on when the container is deleted.

Data volume Features:

• The data volume is copied to the newly initialized data volume when the container initiates initialization, if the image used by the container contains data at the mount point.

• Data volumes can be shared and reused directly in containers.

• You can modify the contents of the data volume directly.

• Changes to the data volume do not affect the update of the mirror.

• The volume will persist even if the container on which the data volume is mounted has been deleted.

1. Data Volume usage

To create and mount a data volume:

$ sudo docker run-itd--name ubuntu_test1-v/container_data:/data Ubuntu

Note: Container_data is a host directory,/data is a directory in the container, directory does not exist automatically created

$ sudo docker inspect Ubuntu_test1

"Mounts": [

{

"Source": "/container_data",

"Destination": "/data",

"Mode": "",

"RW": True

}

],

You can see that it has been mounted successfully and that the container has read and write permissions to the directory.

Test:

$ CD Container_data

$ sudo touch test.txt

$ sudo docker exec ubuntu_test1 ls/data

Test.txt

Files created in the host directory are also seen inside the container.

2. Delete the container, will the data be deleted together?

$ sudo docker stop Ubuntu_test1

$ sudo docker rm ubuntu_test1

$ ls Container_data

Test.txt

The files in the data volume directory on the host are not changed, indicating that deleting the container does not affect the data volume.

3. Reboot a container and mount the data volume again try

$ sudo docker run-itd--name ubuntu_test2-v/container_data:/data Ubuntu

$ sudo docker exec ubuntu_test1 ls/data

Test.txt

The file still exists, indicating that the data was copied to the container during initialization.

3. Start a container and mount the data volume here

$ sudo docker run-itd--name ubuntu_test3-v/container_data:/data Ubuntu

$ sudo docker exec ubuntu_test1 ls/data

Test.txt

You can also see the data that the data volume can share with multiple containers.

Ii. container data volumes

Use a normal container as a data volume, and let other containers implement data sharing by mounting the container.

Note: Data volume containers can degrade I/O performance.

1. Container Data Volume usage

To create a Dvdata data volume container:

$ sudo docker run-itd-v/data--name dvdata Ubuntu

Note:/data is a shared directory within the data volume container

Mount the data volume of the Dvdata container in another container:

$ sudo docker run-itd--name web1--volumes-from dbdata Ubuntu

$ sudo docker run-itd--name web2--volumes-from dbdata ubuntu

Respectively into the WEB1, WEB2 container, there will be a/data directory, in the Web1/data directory to create files, WEB2 can also see.

Summary: If you delete Dvdata, Web1, WEB2, the data volume is not automatically deleted. Use the Docker rm-v command to specify the container to delete if you want to delete the last container to which it is attached.

Blog Address: http://lizhenliang.blog.51cto.com

Third, commit command use

The commit command is the ability to submit the image and modifications in an existing container as a new image, which also preserves the read-write layer content.

1. Start a new container and create a test.txt file in the container/opt directory

$ sudo docker run-itd--name web Ubuntu

$ sudo docker exec web Touch/opt/test.txt

$ sudo docker exec web ls/opt

Test.txt

2. Submit a new image

$ sudo docker commit web Web:v2

$ sudo docker images

650) this.width=650; "src=" Http://s3.51cto.com/wyfs02/M02/79/13/wKiom1aHmizRJTH3AAALZBBhRf0914.png "title=" QQ picture 20160102173631.png "alt=" Wkiom1ahmizrjth3aaalzbbhrf0914.png "/>

3. Successfully submit a new image, use this new image to start a container to see if the created file exists

$ sudo docker run-itd--name web_v2 web:v2

$ sudo docker exec web_v2 ls/opt

Test.txt

Summary: The Commit command also enables the storage of read-write data, but is not suitable for data persistence

Iv. Data Volume container backup and restore

Backup:

$ sudo docker run--volumes-from dvdata-v/container_backup:/backup ubuntu tar cvf/backup/backup.tar/data

Note: Create a temporary container, mount the Dvdata container data volume, mount the data volume/container_backup directory to the container/bakcup, perform a backup/data directory in the container to/backup, that is, back up to the host/container_ Backup directory.

Recovery:

#先创建一个数据卷容器

$ sudo docker run-v/data--name dvdata2 Ubuntu

Note: This data volume directory name should be the same as the backup

#再将备份文件恢复到这个数据卷容器

$ sudo docker run--volumes-from dvdata2-v/container_backup:/backup ubuntu tar Xvf/backup/backup.tar

#最后启动一个容器挂载验证/data Directory Data Recovery succeeded

$ sudo docker run-itd--volumes-from dvdata2--name web_recover Ubuntu

V. Migrating containers and Mirrors

Export and Import commands use:

#export导出容器会丢失历史记录和元数据, similar to the snapshot

Create a Test Container first:

$ sudo docker exec web Touch/opt/test.txt

$ sudo docker exec web ls/opt

Test.txt

To perform an export:

$ sudo docker export web > Web.tar

To perform an import:

$ Cat Web.tar | sudo docker import-web:v2

$ sudo docker images

650) this.width=650; "src=" Http://s2.51cto.com/wyfs02/M00/79/13/wKiom1aHmtuhyu9UAAAHiNrmLo0553.png "title=" QQ picture 20160102173920.png "alt=" Wkiom1ahmtuhyu9uaaahinrmlo0553.png "/>$ sudo docker run-itd--name web_v2 web:v2/ Bin/bash

#启动这个镜像要加/bin/bash, otherwise error error response from Daemon:no command specified

$ sudo docker exec web_v2 ls/opt

Test.txt

Summary: The Export command also allows you to save the data in the container and migrate to another Docker host

The Save and Load commands use:

#一般用于迁移镜像到别处

Export:

$ sudo docker save web > Web.tar

Import:

$ sudo docker load < Ubuntu.tar

Note: History and metadata are not discarded, and versions can be rolled back. Start without adding/bin/bash

This article is from the "Li Zhenliang Technology Blog" blog, make sure to keep this source http://lizhenliang.blog.51cto.com/7876557/1730892

Docker data persistence and container migration