Git – Capistrano and several SSH keys

capistranogitssh

I need Capistrano to use 2 different SSH keys. One is for the git repository, one is for the server to deploy to.

Whichever key I rename to id_rsa in my .ssh folder, works. The other one doesn't. If I rename the git key to id_rsa, Capistrano can connect to the git repository, but then can't authenticate at the server to deploy. If I call it something else, it will not be able to connect to the git repo. I know that the other key works, cause I can do ssh -i ~/.ssh/otherKey.pem and it will successfully connect to the server.

This is what I have in my deploy.rb Capistrano file.

ssh_options[:keys] = [
        File.join(ENV["HOME"], ".ssh", "id_rsa"),
        File.join(ENV["HOME"], ".ssh", "deploy")
    ]

ssh_options[:forward_agent] = true

How can I tell Capistrano to use BOTH the keys? It only seems to use the one called id_rsa.

edit:

Here's the output from Capistrano with the error message:

$ cap yii deploy
  * executing `yii'
Yii
  * executing `deploy'
  * executing `deploy:update'
 ** transaction: start
  * executing `deploy:update_code'
    executing locally: "git ls-remote git@project.beanstalkapp.com:/projectyii.git HEAD"
  * executing "git clone -q git@project.beanstalkapp.com:/projectyii.git /var/www/projectyii-trunk/releases/20110824174629 && cd /var/www/projectyii-trunk/releases/20110824174629 && git checkout -q -b deploy 5e14521285ca04a605353e97bdf31c3a2889dbfb && (echo 5e14521285ca04a605353e97bdf31c3a2889dbfb > /var/www/projectyii-trunk/releases/20110824174629/REVISION)"
    servers: ["yii.project.com"]
    [yii.project.com] executing command
 ** [yii.project.com :: err] Error reading response length from authentication socket.
 ** [yii.project.com :: err] Permission denied (publickey,keyboard-interactive).
 ** [yii.project.com :: err] fatal: The remote end hung up unexpectedly
    command finished
*** [deploy:update_code] rolling back
  * executing "rm -rf /var/www/projectyii-trunk/releases/20110824174629; true"
    servers: ["yii.project.com"]
    [yii.project.com] executing command
    command finished
failed: "sh -c \"git clone -q git@project.beanstalkapp.com:/projectyii.git /var/www/projectyii-trunk/releases/20110824174629 && cd /var/www/projectyii-trunk/releases/20110824174629 && git checkout -q -b deploy 5e14521285ca04a605353e97bdf31c3a2889dbfb && (echo 5e14521285ca04a605353e97bdf31c3a2889dbfb > /var/www/projectyii-trunk/releases/20110824174629/REVISION)\"" on yii.project.com

edit:

Another thing: it totally works fine from my local machine, just not on the deploy server – with exactly the same config files! It seems Capistrano uses the correct keys on my local machine, but not on the deploy machine.

Best Answer

Disclaimer: I don't know anything about Capistrano.

If it simply does normal ssh calls (or calls git to do this), you can configure the right key to use in your ~/.ssh/config on a per-host (or per-host-alias) basis.

For example, I have these lines in my ~/.ssh/config file:

# Git bei Github
Host github.com
User git
IdentityFile ~/.ssh/svn_id_rsa

#  Andere Mathe-Hosts
Host *.math.hu-berlin.de
User ebermann
IdentityFile ~/.ssh/id_rsa
ControlMaster auto

Related Solutions

Git – the difference between ‘git pull’ and ‘git fetch’

In the simplest terms, git pull does a git fetch followed by a git merge.

You can do a git fetch at any time to update your remote-tracking branches under refs/remotes/<remote>/. This operation never changes any of your own local branches under refs/heads, and is safe to do without changing your working copy. I have even heard of people running git fetch periodically in a cron job in the background (although I wouldn't recommend doing this).

A git pull is what you would do to bring a local branch up-to-date with its remote version, while also updating your other remote-tracking branches.

From the Git documentation for git pull:

In its default mode, git pull is shorthand for git fetch followed by git merge FETCH_HEAD.

Git – Difference between “git add -A” and “git add .”

This answer only applies to Git version 1.x. For Git version 2.x, see other answers.

Summary:

git add -A stages all changes
git add . stages new files and modifications, without deletions (on the current directory and its subdirectories).
git add -u stages modifications and deletions, without new files

Detail:

git add -A is equivalent to git add .; git add -u.

The important point about git add . is that it looks at the working tree and adds all those paths to the staged changes if they are either changed or are new and not ignored, it does not stage any 'rm' actions.

git add -u looks at all the already tracked files and stages the changes to those files if they are different or if they have been removed. It does not add any new files, it only stages changes to already tracked files.

git add -A is a handy shortcut for doing both of those.

You can test the differences out with something like this (note that for Git version 2.x your output for git add . git status will be different):

git init
echo Change me > change-me
echo Delete me > delete-me
git add change-me delete-me
git commit -m initial

echo OK >> change-me
rm delete-me
echo Add me > add-me

git status
# Changed but not updated:
#   modified:   change-me
#   deleted:    delete-me
# Untracked files:
#   add-me

git add .
git status

# Changes to be committed:
#   new file:   add-me
#   modified:   change-me
# Changed but not updated:
#   deleted:    delete-me

git reset

git add -u
git status

# Changes to be committed:
#   modified:   change-me
#   deleted:    delete-me
# Untracked files:
#   add-me

git reset

git add -A
git status

# Changes to be committed:
#   new file:   add-me
#   modified:   change-me
#   deleted:    delete-me

Best Answer

Related Solutions

Git – the difference between ‘git pull’ and ‘git fetch’

Git – Difference between “git add -A” and “git add .”

Related Topic