原文地址:https://hub.docker.com/r/cwspear/docker-local-persist-volume-plugin/

Short Description

Create named local volumes that persist in the location(s) you want.
 

Full Description

Local Persist Volume Plugin for Docker

 

Create named local volumes that persist in the location(s) you want!

Rationale

In Docker 1.9, they added support for creating standalone named Volumes. Now with Docker 1.10 and Docker Compose 1.6's new syntax, you can create named volumes through Docker Compose.

This is great for creating standalone volumes and easily connecting them to different directories in different containers as a way to share data between multiple containers. On a much larger scale, it also allows for the use of Docker Volume Plugins to do cool things like Flocker is doing (help run stateful containers across multiple hosts).

Even if something like Flocker is overkill for your needs, it can still be useful to have persistent data on your host. I'm a strong advocate for "Docker for small projects" and not just huge, scaling behemoths and microservices. I wrote this out of a need on projects I'm currently working on and have in production.

This local-persist approach gives you the same benefits of standalone Volumes that docker volume create ... normally affords, while also allowing you to create Volumes that persist, thus giving those stateful containers their state. Read below how to install and use, then read more about the benefits of this approach.

Installing & Running

To create a Docker Plugin, one must create a Unix socket that Docker will look for when you use the plugin and then it listens for commands from Docker and runs the corresponding code when necessary.

Running the code in this project with create the said socket, listening for commands from Docker to create the necessary Volumes.

According to the Docker Plugin API Docs:

Plugins can run inside or outside containers. Currently running them outside containers is recommended.

It doesn't really say why one way is recommended over the other, but I provide binaries and instructions to run outside of container, as well as an image and instructions to run it inside a container.

Running Outside a Container

Quick Way

I provide an install script that will download the proper binary, set up an Systemd service to start when Docker does and enable it.

curl -fsSL https://raw.githubusercontent.com/CWSpear/local-persist/master/scripts/install.sh | sudo bash

This needs be to run on the Docker host. i.e. running that on a Mac won't work (and it will print a message saying as much and exit).

This has been tested on Ubuntu 15.10, and is known not to work on CoreOS (yet). If you need to use Upstart instead of Systemd, you can pass the --upstart flag to the install script, but it isn't as tested, so it may not work:

curl -fsSL https://raw.githubusercontent.com/CWSpear/local-persist/master/scripts/install.sh | sudo bash -s -- --upstart

Follow the same process to update to the latest version.

Manual Way

If you're uncomfortable running a script you downloaded off the internet with sudo, you can extract any of the steps out of the install.sh script and run them manually. However you want to do it, the main steps are:

  1. Download the appropriate binary from the Releases page for your OS and architecture.
  2. Rename the downloaded file docker-volume-local-persist
  3. Place it in /usr/bin (you can put it somewhere else, but be sure your Systemd (or similar) config reflects the change).
  4. Make sure the file is executable (chmod +x /usr/bin/docker-volume-local-persist)
  5. It's enough to just run it at this point (type docker-volume-local-persist and hit enter) to test, etc, and if that's all you're trying to do, you're done. But if you want it to start with Docker, proceed to step 6.
  6. Download systemd.service
  7. Rename the service file to docker-volume-local-persist.service
  8. Move it to /etc/systemd/system/
  9. run sudo systemctl daemon-reload to reload the config
  10. run sudo systemctl enable docker-volume-local-persist to enable the service (it will start after Docker does)
  11. run sudo systemctl start docker-volume-local-persist to start it now. Safe to run if it's already started

Running from Within a Container

I maintain an image on Docker Hub to run this plugin from a container:

docker run -d \
-v /run/docker/plugins/:/run/docker/plugins/ \
-v /path/to/store/json/for/restart/:/var/lib/docker/plugin-data/ \
-v /path/to/where/you/want/data/volume/:/path/to/where/you/want/data/volume/ \
cwspear/docker-local-persist-volume-plugin

The -v /run/docker/plugins/:/run/docker/plugins/ part will make sure the sock file gets created at the right place. You also need to add one or more volumes to places you want to mount your volumes later at.

For example, if I am going to persist my MySQL data for a container I'm going to build later at /data/mysql/, I would add a -v /data/mysql/:/data/mysql/ to the command above (or even -v /data/:/data/). You can add more than one location in this manner.

Lastly, the -v /path/to/store/json/for/restart/:/var/lib/docker/plugin-data/ part is so that the plugin can create a json file to know what volumes existed in case of a system restart, etc.

When the container is destroyed, etc, it will look at a file it created in /var/lib/docker/plugin-data/ to recreate any volumes that had previously existed, so you want that JSON file to persist on the host.

Usage: Creating Volumes

Then to use, you can create a volume with this plugin (this example will be for a shared folder for images):

docker volume create -d local-persist -o mountpoint=/data/images --name=images

Then if you create a container, you can connect it to this Volume:

docker run -d -v images:/path/to/images/on/one/ one
docker run -d -v images:/path/to/images/on/two/ two
# etc

Also, see docker-compose.example.yml for an example to do something like this with Docker Compose (needs Compose 1.6+ which needs Engine 1.10+).

Benefits

This has a few advantages over the (default) local driver that comes with Docker, because our data will not be deleted when the Volume is removed. The local driver deletes all data when it's removed. With the local-persist driver, if you remove the driver, and then recreate it later with the same command above, any volume that was added to that volume will still be there.

You may have noticed that you could do this with data-only containers, too. And that's true, and using that technique has a few advantages, one thing it (specifically as a limitation of volumes-from) does not allow, is mounting that shared volume to a different path inside your containers. Trying to recreate the above example, each container would have to store images in the same directory in their containers, instead of separate ones which local-persist allows.

Also, using local-persist instead of data-only containers, docker ps -a won't have extra dead entries, and docker volume ls will have more descriptive output (because volumes have names).

(转)Docker volume plugin - enabled create local volume on docker host的更多相关文章

  1. [Docker] Create a Volume

    We can create volumn to keep the data, even we stop the container and restart again, the data won't ...

  2. centos7下安装docker(13.1docker存储--data volume)

    我们现在知道docker 有两种存储方式:storage driver和data volume stroage driver这种存储方式主要是存储那些无状态的数据,是镜像层和容器层组成的,而data ...

  3. kubernetes支持local volume

    目录 local volume 创建一个storage class 静态创建PV 使用local volume PV 动态创建PV local volume kubernetes从1.10版本开始支持 ...

  4. k8s local volume 和host path volume的区别

    k8s提供多种volume接口,其中local 和host path是容易混淆的两个接口.下面这篇文章解释了两者的区别: https://groups.google.com/forum/#!topic ...

  5. Docker(三)-Docker中Image、Container与Volume的迁移

    Image 镜像的迁移,适用于离线环境. 一般离线环境,都会自建Docker Registry. 无论 官方的 ,还是最近流行的 Harbor ,都是不错的选择. 但是,这个世界上就是有些环境,或者说 ...

  6. 详解Docker中Image、Container与 Volume 的迁移

    开源Linux 长按二维码加关注~ 上一篇:Linux Used内存到底哪里去了? 已经部署的容器化服务,也不是不需要维护的.而且,由于生产环境往往有这样那样的严格要求,往往需要些非常规操作.Imag ...

  7. Docker源码分析(四):Docker Daemon之NewDaemon实现

    1. 前言 Docker的生态系统日趋完善,开发者群体也在日趋庞大,这让业界对Docker持续抱有极其乐观的态度.如今,对于广大开发者而言,使用Docker这项技术已然不是门槛,享受Docker带来的 ...

  8. [置顶] Docker学习总结(7)——云端基于Docker的微服务与持续交付实践

    本文根据[2016 全球运维大会•深圳站]现场演讲嘉宾分享内容整理而成 讲师简介 易立 毕业于北京大学,获得学士学位和硕士学位:目前负责阿里云容器技术相关的产品的研发工作. 加入阿里之前,曾在IBM中 ...

  9. Docker进阶:容器卷、DockerFile、Docker网络原理

    1.Docker镜像 1.1镜像是什么 镜像是一种轻量级.可执行的独立软件包,用来打包软件运行环境和机遇运行环境开发的软件. 包含一个软件的所有内容.蒋所有的应用和环境,直接打包为docker镜像,直 ...

随机推荐

  1. mysql数据库的笔记

    增删改查置顶: 插入数据: 基本语法 : insert into [表名](字段名1,字段名2……) values(记录1),(记录2): insert into [表名] values(记录1),( ...

  2. Linux:减号(-)详解

    减号(-) 代表标准输出/标准输入, 视命令而定. “-”代替stdin和stdout的用法 为应用程序指定参数 ps -aux tar -zxf test.tar 一个减号和两个减号 一个减号后面跟 ...

  3. [Scala]Scala学习笔记五 Object

    1. 单例对象 Scala没有静态方法或静态字段,可以使用object来达到这个目的,对象定义了某个类的单个实例: object Account{ private var lastNumber = 0 ...

  4. Vim技能修炼教程(7) - 可视模式

    可视模式 可视模式是与正常模式.插入模式一起并列的模式.它的作用就像图形化编辑器下用鼠标来选择一个块. 在vim下,使用正常模式和ex命令,连搜带跳行的,未必就比用鼠标慢. 我们先做一个例子找找感觉, ...

  5. [转] 用深度学习(CNN RNN Attention)解决大规模文本分类问题 - 综述和实践

    转自知乎上看到的一篇很棒的文章:用深度学习(CNN RNN Attention)解决大规模文本分类问题 - 综述和实践 近来在同时做一个应用深度学习解决淘宝商品的类目预测问题的项目,恰好硕士毕业时论文 ...

  6. webpy/flask/bottle性能测试

    这三个都是Python WSGI的web开发框架,到底用哪个呢?单纯从性能的角度而言,可能哪个快就用哪个,但是这也不是绝对的.比如我就比较喜欢webpy的router配置放在一个文件中,而flask和 ...

  7. 02-C与OC语言的一些小知识

    1.        #import 跟#include.@class有什么区别?#import<> 跟 #import”"又什么区别? 1>  #import和#inclu ...

  8. weight decay (权值衰减)

    http://blog.sina.com.cn/s/blog_890c6aa30100z7su.html 在机器学习或者模式识别中,会出现overfitting,而当网络逐渐overfitting时网 ...

  9. .pyc和.pyo文件有何用

    百度知道:http://zhidao.baidu.com/link?url=_tFP1xglFnoEBObWtIArI3b3Ft0PQowx5m5ruIaX3mFIAFVr7vX45Lfb0geCjA ...

  10. debian 8.1 安装idempiere 2.1 X64 笔记

    接上文.当虚拟服务器和虚拟机搭建完成后.登陆debian 8.1 X64. 进入虚拟服务器控制台.打开虚拟机.root登陆.(留好初始状态系统快照.以便系统恢复.) 由于之前debian8.1X64默 ...