如何为Git存储库组织和设置镜像备份服务器？ [英] How to organize and setup mirrored backup servers for Git repositories?

问题描述

我正在将一些svn版本库移到Git中。所以，我基本上试图做的是这样的：

• 设置一台带有Git仓库的服务器，我将从中抽取并推送到 li>
  
• 为第一台服务器上的所有存储库设置一些备份服务器。
，假设我在我的服务器上有一个目录，例如： $ HOME / git / ，它只有裸露的存储库。例如：



 〜/ git / project1.git 
〜/ git / project2.git 
〜/ git / project3.git 
 ... 
  

我的备份服务器可能是此服务器的镜像，或将备份的数据保存在档案或其他内容中。我想我可以这样做：

  git clone --bare ssh：// gitserver /〜user / git / projectX。 git 
  

或者：



  $ cd〜/ git / project1.git 
 $ git bundle create〜/ gitbackup / project1.bdl --all 
   
但是，拥有大量项目或者任何策略都是一项乏味的任务，因此在每种情况下，我都需要制作一些脚本来自动执行任务。
 
 
 
我想知道你们是怎么做的？也许有更好的办法来做到这一点，而不是我已经考虑过的。任何提示将不胜感激。
解决方案
总体思路是： 
 
  
 
在您的备份服务器上创建裸回购
 
 
有钩子（可能是接收后挂钩），然后通过  git：协议或 ssh + git：协议。
 
 

还有其他技巧早在2008年，基于 gibak ，但这个想法保持不变。

---

后接收钩子的示例：

• this one from Grant Limberg ：

 
＃！/ bin / sh 
＃
＃在成功完成
＃commit后调用。 
＃
＃把这个文件放在.git / hooks和chmod + x 
 
 BRANCH =`git branch --no-color 2> / dev / null | sed -e'/ ^ [^ *] / d'-e's / * \（。* \）/ \ 1 /'`
 git push origin $ BRANCH 
 

• that one （in ruby​​）

 
＃！/ usr / bin / env ruby​​ 
 STDIN.read.split（\\\
）。 
 oldrev，newrev，refname = line.split（''）
 
如果refname.match（/ ^ refs \ / heads\ /(.*)/）
分支= $ 1 
 
`git push origin＃{branch}`
 else 
 puts＃{refname}很奇怪，不知道该怎么办。 
结束
结束
 

I am moving some svn repositories to Git. So, what I basically try to do is this:

• Setup one server with bare Git repositories from which I will pull and push to
• Setup a few backup servers for all of my repositories that are on the 1st server.

So, let's say i have a directory on my server, like: $HOME/git/, which has bare repositories. Eg:

~/git/project1.git
~/git/project2.git
~/git/project3.git
...

My backup servers may be mirrors to this server, or keep the backed up data in archives or whatever. I suppose I can do something like:

git clone --bare ssh://gitserver/~user/git/projectX.git

Or maybe:

$ cd ~/git/project1.git
$ git bundle create ~/gitbackup/project1.bdl --all

and then copy all bundles from all projects to my backup servers.
However, having a lot of projects either strategy would be a tedious task, so in each case I would need to make some scripts to automate the task.

I wonder how are you guys doing this? Maybe there is some better way to do it than what I considered already. Any tip would be appreciated.

解决方案

The general idea would be to:

• create bare repo on your backup server
• have hooks (probably a post-receive hook) on your main bare Git repository to automatically push whatever is received to the corresponding Git backup repo, either through git: protocol or ssh+git: protocol.

There are other techniques back in 2008, based on gibak, but the idea remains the same.

---

Example of a post-receive hook:

• this one from Grant Limberg:

    #!/bin/sh
    #
    # A hook script that is called after a successful
    # commit is made.
    #
    # Place this file in .git/hooks and chmod +x

    BRANCH=`git branch --no-color 2> /dev/null | sed -e '/^[^*]/d' -e 's/* \(.*\)/\1/'`
    git push origin $BRANCH

• that one (in ruby)

    #!/usr/bin/env ruby
    STDIN.read.split("\n").each do |line|
       oldrev, newrev, refname = line.split(' ')

       if refname.match(/^refs\/heads\/(.*)/)
         branch = $1

         `git push origin #{branch}`
       else
         puts "#{refname} was weird, not sure what to do."
       end
    end

这篇关于如何为Git存储库组织和设置镜像备份服务器？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！