Shell Configuration Deployer (scd)

A program used to deploy shell configuration to remote hosts. The program can be used to automatically deploy your shell configuration to remote hosts before sshing in to them. This way you can always use your desired shell configuration no matter where you are.

scd keeps track of which servers have correct shell configuration by keeping track of the time of deployment as well as a list of programs and files that have been installed. Any files that have since changed or been added will be redeployed to the server. It can not handle removal of files or programs.

Usage

Use scd on the command line to deploy your configuration to the desired host:

scd -u vagrant -p vagrant -P 2222 127.0.0.1

Use ~/.scd/config to specify what programs should be installed on the remote server and what files should be deployed. Example:

{
    "user": "vagrant",
    "host": "127.0.0.1",
    "port": 2222,
    "shell": "zsh",
    "private_key": "~/my_key.pem",
    "ignored_files": [
        "*/.gitignore",
        "*/.git/*",
        "*/.DS_Store"
    ],
    "files": [
        "~/.oh-my-zsh",
        "~/.zshrc",
        "~/.gitconfig",
        ["~/my_settings.txt", "~/server_settings.txt"]
    ],
    "programs": [
        "tree"
    ],
     "scripts": [
         {
            "file": "~/init.sh",
            "as_sudo": false
        }
     ]
}

This configuration will deploy the folder .oh-my-zsh and the files .zshrc and .gitconfig located in the users home folder and place them in /home/<user> on to the remote host. my_settings.txt will be deployed as server_settings.txt. Any .gitignore and .DS_Store files will be ignored as well as .git folders.

It will also install tree and zshand set zsh as the default login shell for the user and run the script ~/init.sh on the remote host.

If a password is provided, scd will run commands on the remote host as sudo. This is accomplished by sending a file containing the password to the host and executing the commands using cat password_file | sudo -S <command>. This ensures that the password is never readable from other processes. The file is removed directly after executing the script, even on failure.

Installation

• Make sure you have python3 installed.
• Clone the repository

git clone git@github.com:timlindeberg/ShellConfigDeployer.git <SHELL_CONFIG_DEPLOYER>

• Install scd:

cd <SHELL_CONFIG_DEPLOYER>
python3 setup.py install

• You can now use scd on the command line. When running it for the first time a default configuration will be created.

• Edit the configuration file (~/.scd/config) with your settings

Configuration options

"shell"

Specifies which shell to use on the remote server. If specified, the selected shell will be installed together with the other specified programs and the default login shell will be changed to this shell.

"programs"

Specifies a list of programs to install on the remote host.

"files"

Specifies a list of files and directories to deploy to the remote host. The paths specified are absolute and will be deployed to the same location on the host. ~ will expand to the home folder on the local host and to /home/<USER> on the remote host. You can also specify the name of the file or folder on the remote host using either a dictionary or a list:

{
    ...
    "files": [ 
        "~/.oh-my-zsh",
        ["~/a.txt", "~/b.txt"],
        {
            "source_path": "/a/b/c.txt",
            "host_path": "/c/b/a.txt"
        }
    ]
    ...
}

This configuration will deploy the folder ~/.oh-my-zsh to /home/<USER>/.oh-my-zsh on the remote host, ~/a.txt to /home/<USER>/b.txt and /a/b/c.txt to /c/b/a.txt.

"scripts"

A list of scripts to execute on the remote host. Each entry should be an object containg the entries "file", a path to the script to be executed, and "as_sudo", a boolean determining whether to execute the script as sudo or not.

If a script executes successfully it won't be executed again. Since the scripts are executed on the remote host they cannot access local files or environment variables and will instead be executed in the remote environment. If a script exits with an exit code other than 0 the execution will be considered a failure and scd will try to execute the script again on the subsequent run. Scripts should preferably be idempotent.

"ignored_files"

Specifies a list of files and directories that will be ignored when looking for files to deploy. The ignored files follow the same style as .gitignore. For example */.git/* will ignore all files inside .git folders and */.DS_Store will ignore all .DS_Storefiles anywhere.

"user"

Selects which user to authenticate against the host with. Can also be specified with the flag --user (-u) if you use different user names for different hosts. The user needs sudo rights in order to install programs.

"host"

Selects which host to deploy the configuration to. Normally this would be given as a command line argument but if you usually connect to the same host you can specify it in the config file.

"port"

Selects which port to connect through. Can also be specified using the flags --port (-P). Defaults to 22.

"private_key"

Path to a private key to use when connecting to servers. Can also be specified using the flags --private_key (-i). If not specified, no private will be used.

"password"

The password can be stored in the config file though it is not recommended.

Flags

--help

Prints a help message and exits.

--port (-P) PORT

Specifies which port to connect to, defaults to 22.

--private_key (-i) PRIVATE_KEY

Specifies a path to a private key to use when connecting to servers. Similar to the -i flag of the ssh command.

--read-password (-r)

Reads the password from user input. If combined with the --private-key flag the password will be used to decrypt the private key.

--password-file (-f) PATH

Specifies a path to a file containing the password to use when authenticating against the host. This is preferable to passing a password directly using the -p flag since otherwise the password can be seen by other processes. If combined with the --private-key flag the password will be used to decrypt the private key.

--password (-p) PASSWORD

Specify the password to use for authentication. It is better to use the -f flag to specify password since otherwise the password can be seen by other processes. Only use this for testing or if you don't care about security. If combined with the --private-key flag the password will be used to decrypt the private key.

--verbose (-v)

Prints additional information.

--user (-u) USER

Specify which user to authenticate with.

--clear-status HOST

Clears the status information for a given host.

--force

Forces a full redeployment of the configuration, ignoring the status of the host.

--host-status

Prints the status of all hosts and exits.

--print-config

Prints the current configuration and exits.