path: root/README

etckeeper is a collection of tools to let /etc be stored in a git or
mercurial repository. It hooks into apt to automatically commit changes
made to /etc during package upgrades. It uses [metastore][1] to track file
metadata that git does not normally support, but that is important for
/etc, such as the permissions of `/etc/shadow`. It's quite modular and
configurable, while also being simple to use if you understand the basics
of working with version control.

   [1]: http://david.hardeman.nu/software.php

## security warnings

First, a big warning: By checking /etc into revision control, you are
creating a copy of files like /etc/shadow that must remain secret. Anytime
you have a copy of a secret file, it becomes more likely that the file
contents won't remain secret. etckeeper is careful about file permissions,
and will make sure that repositories it sets up don't allow anyone but root
to read their contents. However, you *also* must take care when cloning
or copying these repositories, not to allow anyone else to see the data.

Since git mushes all the files into packs under the .git directory, the
whole .git directory content needs to be kept secret. (Ditto for mercurial
and .hg)

Also, since revision control systems don't keep track of the mode of files
like the shadow file, it will check out world readable, before etckeeper
fixes the permissions. The tutorial has some examples of safe ways to avoid
these problems when cloning an /etc repository.

Also note that `etckeeper init` runs code stored in the repository.
So don't use it on repositories from untrusted sources.


## what etckeeper does

etckeeper has special support to handle changes to /etc caused by
installing and upgrading packages. Before apt installs packages,
`etckeeper pre-install` will check that /etc contains no uncommitted changes.
After apt installs packages, `etckeeper post-install` will add any new
interesting files to the repository, and commit the changes.

Revsion control systems are designed as a way to manage source code, not as
a way to manage arbitrary directories like /etc. This means there are a few
limitations that etckeeper has to work around. These include file metadata
storage, empty directories, and special files.

git and mercurial have only limited tracking of file metadata, being able
to track the executable bit, but not other permissions or owner info. So
file metadata storage is handled by `metastore`. Amoung other chores,
`etckeeper init` sets up a `pre-commit` hook that uses `metastore` to store
metadata about file owners, permissions, and even extended attributes into
a `/etc/.metadata` file. This metadata is stored in git along with
everything else, and can be applied if the repo should need to be checked
back out.

git and mercurial cannot track empty directories, but they can be
significant sometimes in /etc. So the `pre-commit` hook also stores
information that can be used to recreate the empty directories in a
`/etc/.etckeeper` file.

git and mercurial don't support several special files that you _probably_
won't have in /etc, such as unix sockets, named pipes, hardlinked files
(but softlinks are fine), and device files. The `pre-commit` hook will warn
if your /etc contains such special files.


## tutorial

A quick walkthrough of using etckeeper.

First, edit `/etc/etckeeper/etckeeper.conf` to select which version control
system to use. The default is git, and this tutorial assumes you're using
it. Mercurial is similar.

The `etckeeper init` command initialises an /etc/.git/ repository. This
command is careful to never overwrite existing files or directories in
/etc. It will create a `.gitignore` if one doesn't already exist, sets up
pre-commit hooks if they don't already exist, and so on. It does *not*
commit any files, but does `git add` all interesting files for an initial
commit later.

	etckeeper init

Now you might want to run `git status` to check that it includes all
the right files, and none of the wrong files. And you can edit the
`.gitignore` and so forth. Once you're ready, it's time to commit:

	cd /etc
	git commit -m "initial checkin"
	git gc # pack git repo to save a lot of space

After this first commit, you can use regular git commands to handle
further changes:

	passwd someuser
	git status
	git commit -a -m "changed a password"

Rinse, lather, repeat. You might find that some files are changed by
daemons and shouldn't be tracked by git. These can be removed from git:

	git rm --cached printcap # modified by CUPS
	echo printcap >> .gitignore
	git commit -a -m "don't track printcap" 

etckeeper hooks into apt so changes to interesting files in /etc caused by
installing or upgrading packages will automatically be committed. Here
"interesting" means files that are not ignored by `.gitignore`.

You can use any git commands you like, but do keep in mind that, if you
check out a different branch or an old version, git is operating directly
on your system's /etc. But if you do decide to check out a branch or tag,
make sure you run "etckeeper init" again, to get any metadata changes:

	git checkout april_first_joke_etc
	etckeeper init

Often it's better to clone /etc to elsewhere and do potentially dangerous
stuff in a staging directory. You can clone the repository using git clone,
but be careful that the directory it's cloned into starts out mode 700, to
prevent anyone else from seeing files like shadow, before `etckeeper init`
fixes their permissions:

	mkdir /my/workdir
	cd /my/workdir
	chmod 700 .
	git clone /etc
	cd etc
	etckeeper init .
	chmod 755 ..

Another common reason to clone the repository is to make a backup to a
server. When using git push to create a new remote clone, make sure the new
remote clone is mode 700! (And, obviously, only push over a secure
transport like ssh, and only to a server you trust.)

	ssh server 'mkdir /etc-clone; cd /etc-clone; chmod 700 .; git init'
	git push ssh://server/etc-clone master

Of course, it's also possible to pull changes from a server onto client
machines, to deploy changes to /etc. You might even set up branches for
each machine and merge changes between them. Once /etc is under version
control, the sky's the limit..


## configuration

The main configuration file is `/etc/etckeeper/etckeeper.conf`

etckeeper uses `run-parts` to run the executable files in
`/etc/etckeeper/$command.d/`. You can modify these files, or add your own
custom files. Each individual file is short, simple, and does only one
action.

For example, here's how to configure it to run `git gc` after each apt run,
which will save a lot of disk space:

	cd /etc/etckeeper/post-install.d
	(echo '#!/bin/sh' ; echo 'exec git gc') > 99git-gc
	chmod +x 99git-gc
	git add .
	git commit -m "run git gc after each apt run"

Here's how to disable the automatic commits after each apt run, while still
letting it git add new files and git rm removed ones:

	chmod -x /etc/etckeeper/post-install.d/75vcs-commit


## inspiration

Two blog posts provided inspiration for techniques used by etckeeper:
* http://www.jukie.net/~bart/blog/20070312134706
* http://bryan-murdock.blogspot.com/2007/07/put-etc-under-revision-control-with-git.html

[isisetup][2] has some of the same aims as etckeeper, however, unlike it,
etckeeper does not aim to be a git porcelain with its own set of commands
for manipulating the /etc repository. Instead, etckeeper provides a simple
setup procedure and hooks for setting up an /etc repsository, and then gets
out of your way; you manage the repository using regular git commands.

   [2]: http://www.isisetup.ch/


## license

etckeeper is licensed under version 2 or greater of the GNU GPL.


## author

Joey Hess <joey@kitenet.net>