如何从 git 存储库中删除旧历史记录？

How do I remove the old history from a git repository?

提问by ebneter

I'm afraid I couldn't find anything quite like this particular scenario.

恐怕我找不到任何与此特定场景相似的内容。

I have a git repository with a lot of history: 500+ branches, 500+ tags, going back to mid-2007. It contains ~19,500 commits. We'd like to remove all of the history before Jan 1, 2010, to make it smaller and easier to deal with (we would keep a complete copy of the history in an archive repository).

我有一个历史悠久的 git 存储库：500 多个分支，500 多个标签，可以追溯到 2007 年年中。它包含约 19,500 次提交。我们希望在 2010 年 1 月 1 日之前删除所有历史记录，以使其更小且更易于处理（我们将在存档存储库中保留历史记录的完整副本）。

I know the commit that I want to have become the root of the new repository. I can't, however, figure out the correct git mojo to truncate the repo to start with that commit. I'm guessing some variant of

我知道我想要成为新存储库根的提交。但是，我无法找出正确的 git mojo 来截断 repo 以从该提交开始。我猜有一些变种

git filter-branch

involving grafts would be necessary; it might also be necessary to treat each of the 200+ branches we want to keep separately and then patch the repo back together (something I doknow how to do).

涉及移植将是必要的；它也可能是必要的对待每一个我们要分别保持在200多个分支机构，然后修补回购重新走到一起（这是我不知道该怎么做）。

Has anyone ever done something like this? I've got git 1.7.2.3 if that matters.

有没有人做过这样的事情？如果重要的话，我有 git 1.7.2.3。

采纳答案by apenwarr

Just create a graftof the parent of your new root commit to no parent (or to an empty commit, e.g. the real root commit of your repository). E.g. echo "<NEW-ROOT-SHA1>" > .git/info/grafts

只需将您的新根提交的父项移植到无父项（或空提交，例如存储库的真正根提交）。例如echo "<NEW-ROOT-SHA1>" > .git/info/grafts

After creating the graft, it takes effect right away; you should be able to look at git logand see that the unwanted old commits have gone away:

创建嫁接后，立即生效；您应该能够查看git log并看到不需要的旧提交已经消失：

$ echo 4a46bc886318679d8b15e05aea40b83ff6c3bd47 > .git/info/grafts
$ git log --decorate | tail --lines=11
commit cb3da2d4d8c3378919844b29e815bfd5fdc0210c
Author: Your Name <[email protected]>
Date:   Fri May 24 14:04:10 2013 +0200

    Another message

commit 4a46bc886318679d8b15e05aea40b83ff6c3bd47 (grafted)
Author: Your Name <[email protected]>
Date:   Thu May 23 22:27:48 2013 +0200

    Some message

If all looks as intended, you can just do a simple git filter-branch -- --allto make it permanent.

如果一切看起来都符合预期，您可以做一个简单的事情git filter-branch -- --all来使其永久化。

BEWARE:after doing the filter-branchstep, all commit ids will have changed, so anybody using the old repo must never merge with anyone using the new repo.

注意：在执行filter-branch步骤后，所有提交 ID 都将发生变化，因此使用旧仓库的任何人都不得与使用新仓库的任何人合并。

回答by Alexandre T.

Maybe it's too late to post a reply, but as this page is the first Google's result, it may still be helpful.

也许现在回复已经太晚了，但由于这个页面是 Google 的第一个结果，所以它可能仍然有帮助。

If you want to free some space in your git repo, but do not want to rebuild all your commits (rebase or graft), and still be able to push/pull/merge from people who has the full repo, you may use the git cloneshallowclone (--depthparameter).

如果您想释放 git 仓库中的一些空间，但不想重建所有提交（rebase 或嫁接），并且仍然能够从拥有完整仓库的人那里推/拉/合并，您可以使用git克隆浅克隆（--depth参数）。

; Clone the original repo into limitedRepo
git clone file:///path_to/originalRepo limitedRepo --depth=10

; Remove the original repo, to free up some space
rm -rf originalRepo
cd limitedRepo
git remote rm origin

You may be able to shallow your existing repo, by following these steps:

您可以按照以下步骤浅化现有的存储库：

; Shallow to last 5 commits
git rev-parse HEAD~5 > .git/shallow

; Manually remove all other branches, tags and remotes that refers to old commits

; Prune unreachable objects
git fsck --unreachable ; Will show you the list of what will be deleted
git gc --prune=now     ; Will actually delete your data

How to remove all git local tags?

如何删除所有 git 本地标签？

Ps: Older versions of git didn't support clone/push/pull from/to shallow repos.

Ps：旧版本的 git 不支持从/到浅存储库的克隆/推/拉。

回答by Chris Maes

This methodis easy to understand and works fine. The argument to the script ($1) is a reference (tag, hash, ...) to the commit starting from which you want to keep your history.

这种方法很容易理解并且效果很好。脚本 ( $1)的参数是对要保留历史记录的提交的引用（标签、哈希、...）。

#!/bin/bash
git checkout --orphan temp  # create a new branch without parent history
git commit -m "Truncated history" # create a first commit on this branch
git rebase --onto temp  master # now rebase the part of master branch that we want to keep onto this branch
git branch -D temp # delete the temp branch

# The following 2 commands are optional - they keep your git repo in good shape.
git prune --progress # delete all the objects w/o references
git gc --aggressive # aggressively collect garbage; may take a lot of time on large repos

NOTEthat old tags will still remain present; so you might need to remove them manually

请注意，旧标签仍然存在；所以你可能需要手动删除它们

remark:I know this is almost the same aswer as @yoyodin, but there are some important extra commands and informations here. I tried to edit the answer, but since it is a substantial change to @yoyodin's answer, my edit was rejected, so here's the information!

备注：我知道这与@yoyodin 几乎相同，但这里有一些重要的额外命令和信息。我试图编辑答案，但由于这是对@yoyodin 答案的重大更改，我的编辑被拒绝了，所以这里是信息！

回答by yoyodyn

Try this method How to truncate git history:

试试这个方法如何截断git历史：

#!/bin/bash
git checkout --orphan temp 
git commit -m "Truncated history"
git rebase --onto temp  master
git branch -D temp

Here $1is SHA-1 of the commit you want to keep and the script will create new branch that contains all commits between $1and masterand all the older history is dropped. Note that this simple script assumes that you do not have existing branch called temp. Also note that this script does not clear the git data for old history. Run git gc --prune=all && git repack -a -f -F -dafter you've verified that you truly want to lose all history. You may also need rebase --preserve-mergesbut be warned that the git implementation of that feature is not perfect. Inspect the results manually if you use that.

这$1是您要保留的提交的 SHA-1，脚本将创建包含所有提交之间的新分支，$1并且master删除所有旧历史记录。请注意，这个简单的脚本假定您没有名为 的现有分支temp。另请注意，此脚本不会清除旧历史的 git 数据。运行git gc --prune=all && git repack -a -f -F -d您确认后，你真正想失去所有的历史。您可能还需要rebase --preserve-merges但要注意该功能的 git 实现并不完美。如果您使用它，请手动检查结果。

回答by Jeff Bowman

As an alternative to rewriting history, consider using git replaceas in this article from the Pro Gitbook. The example discussed involves replacing a parent commit to simulate the beginning of a tree, while still keeping the full history as a separate branch for safekeeping.

作为重写历史记录的替代方法，请考虑使用Pro Git书籍git replace中的这篇文章。所讨论的示例涉及替换父提交以模拟树的开始，同时仍将完整历史记录作为单独的分支进行保管。

回答by koppor

If you want to keepthe upstreamrepository with full history, but local smaller checkouts, do a shallow clone with git clone --depth=1 [repo].

如果您想保留具有完整历史记录的上游存储库，但要保留本地较小的结帐，请使用.git clone --depth=1 [repo]

After pushing a commit, you can do

推送提交后，您可以执行

1. git fetch --depth=1to prune the old commits. This makes the old commits and their objects unreachable.
2. git reflog expire --expire-unreachable=now --all. To expire all old commits and their objects
3. git gc --aggressive --prune=allto remove the old objects

1. git fetch --depth=1修剪旧的提交。这使得旧的提交及其对象无法访问。
2. git reflog expire --expire-unreachable=now --all. 使所有旧提交及其对象过期
3. git gc --aggressive --prune=all删除旧对象

See also How to remove local git history after a commit?.

另请参阅如何在提交后删除本地 git 历史记录？.

Note that you cannot push this "shallow" repository to somewhere else: "shallow update not allowed". See Remote rejected (shallow update not allowed) after changing Git remote URL. If you want to to that, you have to stick with grafting.

请注意，您不能将此“浅”存储库推送到其他地方：“不允许浅更新”。请参阅更改 Git 远程 URL 后远程被拒绝（不允许浅更新）。如果你想这样做，你必须坚持嫁接。

回答by Maikel

I needed to read several answers and some other info to understand what I was doing.

我需要阅读几个答案和一些其他信息来了解我在做什么。

1. Ignore everything older than a certain commit

1. 忽略比某个提交更早的所有内容

The file .git/info/graftscan define fake parents for a commit. A line with just a commit id, says that the commit doesn't have a parent. If we wanted to say that we care only about the last 2000 commits, we can type:

该文件.git/info/grafts可以为提交定义假父母。只有提交 ID 的一行表示提交没有父级。如果我们想说我们只关心最近的 2000 次提交，我们可以输入：

git rev-parse HEAD~2000 > .git/info/grafts

git rev-parse gives us the commit id of the 2000th parent of the current commit. The above command will overwrite the grafts file if present. Check if it's there first.

git rev-parse 为我们提供了当前提交的第 2000 个父项的提交 ID。如果存在，上述命令将覆盖嫁接文件。首先检查它是否在那里。

2. Rewrite the Git history (optional)

2.重写Git历史（可选）

If you want to make this grafted fake parent a real one, then run:

如果你想让这个嫁接的假父母成为真正的父母，那么运行：

git filter-branch -- --all

It will change all commit ids. Every copy of this repository needs to be updated forcefully.

它将更改所有提交 ID。此存储库的每个副本都需要强制更新。

3. Clean up disk space

3.清理磁盘空间

I didn't done step 2, because I wanted my copy to stay compatible with the upstream. I just wanted to save some disk space. In order to forget all the old commits:

我没有完成第 2 步，因为我希望我的副本与上游保持兼容。我只是想节省一些磁盘空间。为了忘记所有旧的提交：

git prune
git gc

Alternative: shallow copies

替代方案：浅拷贝

If you have a shallow copy of another repository and just want to save some disk space, you can update .git/shallow. But be careful that nothing is pointing at a commit from before. So you could run something like this:

如果您有另一个存储库的浅表副本并且只想节省一些磁盘空间，您可以更新.git/shallow. 但要小心，没有任何东西指向之前的提交。所以你可以运行这样的东西：

git fetch --prune
git rev-parse HEAD~2000 > .git/shallow
git prune
git gc

The entry in shallow works like a graft. But be careful not to use grafts and shallow at the same time. At least, don't have the same entries in there, it will fail.

浅层的进入就像嫁接一样。但注意不要同时使用嫁接和浅层。至少，那里没有相同的条目，它会失败。

If you still have some old references (tags, branches, remote heads) that point to older commits, they won't be cleaned up and you won't save more disk space.

如果您仍有一些指向旧提交的旧引用（标签、分支、远程头），它们将不会被清除，您也不会节省更多磁盘空间。

回答by HMagdy

When rebaseor pushto head/masterthis error may occurred

当rebase或pushto head/master 时可能会发生此错误

remote: GitLab: You are not allowed to access some of the refs!
To git@giturl:main/xyz.git
 ! [remote rejected] master -> master (pre-receive hook declined)
error: failed to push some refs to 'git@giturl:main/xyz.git'

To resolve this issue in git dashboard should remove master branch from "Protected branches"

要在 git 仪表板中解决此问题，应从“受保护的分支”中删除主分支

then you can run this command

然后你可以运行这个命令

git push -f origin master

or

或者

git rebase --onto temp  master

回答by Shital Shah

There are too many answers here which are not current and some don't fully explain the consequences. Here's what worked for me for trimming down the history using latest git 2.26:

这里有太多不是最新的答案，有些没有完全解释后果。以下是使用最新的 git 2.26 修剪历史记录对我有用的方法：

First create a dummy commit. This commit will appear as the first commit in your truncated repo. You need this because this commit will hold all base files for the history you are keeping. The SHA is the ID of the previous commitof the commit you want to keep (in this example, 8365366). The string 'Initial' will show up as commit message of the first commit. If you are using Windows, type below command from Git Bash command prompt.

首先创建一个虚拟提交。此提交将显示为您截断的存储库中的第一个提交。您需要这个，因为此提交将保存您保留的历史记录的所有基本文件。SHA 是您要保留的提交的前一个提交的 ID （在本例中为8365366）。字符串 'Initial' 将显示为第一次提交的提交消息。如果您使用的是 Windows，请从 Git Bash 命令提示符键入以下命令。

# 8365366 is id of parent commit after which you want to preserve history
echo 'Initial' | git commit-tree 8365366^{tree}

Above command will print SHA, for example, d10f7503bc1ec9d367da15b540887730db862023.

上面的命令将打印 SHA，例如，d10f7503bc1ec9d367da15b540887730db862023.

Now just type:

现在只需输入：

# d10f750 is commit ID from previous command
git rebase --onto d10f750 8365366

This will first put all files as-of commit 8365366in to the dummy commit d10f750. Then it will play back all commits after8365366 over the top of d10f750. Finally masterbranch pointer will be updated to last commit played back.

这将首先将所有文件提交8365366到虚拟提交中d10f750。然后它将播放8365366之后的所有提交d10f750。最后master分支指针将更新到最后一次提交回放。

Now if you want to push these truncated repo, just do git push -f.

现在，如果您想推送这些截断的 repo，只需执行git push -f.

Few things to keep in mind (these applies to other methods as well as this one): Tags are not transferred. While commit IDs and timestamps are preserved, you will see GitHub show these commits in lumpsum heading like Commits on XY date.

需要记住的几件事（这些也适用于其他方法以及本方法）： 标签不会被转移。在保留提交 ID 和时间戳的同时，您将看到 GitHub 以总标题形式显示这些提交，例如Commits on XY date.

Fortunately it is possible to keep truncated history as "archive" and later you can join back trimmed repo with archive repo. For doing this, see this guide.

幸运的是，可以将截断的历史记录保留为“存档”，稍后您可以将修剪过的 repo 与存档 repo 结合起来。为此，请参阅本指南。

回答by RahulMohan Kolakandy

you can delete the directory, files and also the entire history related to the dir or file using the below mentioned jar [ download it] and the commands

您可以使用下面提到的 jar [下载] 和命令删除目录、文件以及与目录或文件相关的整个历史记录

bfg.jar file: https://rtyley.github.io/bfg-repo-cleaner/

bfg.jar 文件：https://rtyley.github.io/bfg-repo-cleaner/

git clone --bare repo-url cd repo_dir java -jar bfg.jar --delete-folders folder_name git reflog expire --expire=now --all && git gc --prune=now --aggressive git push --mirror repo_url

git clone --bare repo-url cd repo_dir java -jar bfg.jar --delete-folders folder_name git reflog expire --expire=now --all && git gc --prune=now --aggressive git push --mirror repo_url