Hap - A simple and effective provisioner.

Hap helps manage build scripts with git and run them concurrently on multiple remote hosts using composable blocks.

First, hap create to setup a new local repo. Then add hosts to the generated Hapfile. Once hosts are in place, run hap build to execute the build blocks and commands specified in the Hapfile for each host. After hap build a .happened file is saved with the current sha of remote repo. To run hap build again a new commit is required or use the --force param.

To run arbitrary commands use hap c, and to execute individual scripts with hap exec. hap c will not push the latest or run from the current directory or use the environment files, those must be added as part of the command. hap exec will push the latest, use the current directory and the environment.

If you only have one host, just use the default section. Then the -h,--host flag while running hap is not necessary.

Make sure every build script is executable before committing to the local repo.

Installation

via Go

go get -u github.com/gwoo/hap/cmd/hap

Binaries

darwin/amd64

curl -L -C - -o /usr/local/bin/hap https://github.com/gwoo/hap/releases/download/v2.6/hap-darwin-amd64; chmod a+x /usr/local/bin/hap

linux/amd64

curl -L -C - -o /usr/local/bin/hap https://github.com/gwoo/hap/releases/download/v2.6/hap-linux-amd64; chmod a+x /usr/local/bin/hap

Basic Workflow

• Run hap create <name>
• Modify Hapfile
• Run hap -h <host> build

Environment Variables

Hap exports HAP_HOSTNAME, HAP_USER, HAP_ADDR, HAP_IP, HAP_PORT for use in scripts. You can add your own by using the env section or the env statement in the host section.

Hapfile

The Hapfile uses git-config syntax. There are 5 sections, default, host, build, include, and env. The default section holds host config that will be applied to all hosts. The host section holds a named host config. A host config includes addr, username, password, identity, build, and cmd, env. Only addr is required. The identity should point to a local ssh private key that has access to the host via the authorized_keys. The build section holds mulitple cmds that could be applied to a host. Multiple build, cmd, or env are permitted for each host. In addition, an include section accepts multiple path statements and an env section accepts multiple file statements.

sections

• host: Holds the configuration for a machine
  • addr: the host:port of the remote machine
  • dir: base directory on the remote machine
  • username: the name of the user to login and run commands
  • password: password for ssh password based authentication
  • identity: path to ssh private key for key based authentication
  • build: one or more groups of commands to run
  • cmd: one or more commands to run on a specific host
  • env: one or more environment files to apply to this host (can override env sections)
• deploy: Holds the configuration for a deploy
  • host: one or more hosts
  • build: one or more groups of commands to run
  • cmd: one or more commands to run on a specific host
  • env: one or more environment files to apply to this host (can override env sections)
• build: sets of commands to run
  • cmd: one or more commands to run
• default : Holds the standard configurations that can be applied to all hosts
• include: Allows other files to be included in the current configuration
  • path: a path to the Hapfile the hap
• env: make variables available to the all commands
  • file: path to a file that can be sourced

Example Hapfile

A default build is specified, so init.sh and update.sh are executed for each host. Host one specifies two commands, notify.sh and cleanup.sh, to be run after the default build commands. For host one, the HAP_HOSTNAME will be one, the HAP_USER will be root, and the HAP_ADDR will be 10.0.20.10:22. Host two specifies no commands, so only the default build will be applied. For host two, the HAP_HOSTNAME will be two, the HAP_USER will be admin, and the HAP_ADDR will be 10.0.20.11:22. There is also a deploy that will return the hostname for one and two.

[default]
  username = "root"
  identity = "~/.ssh/id_rsa"
  build = "initialize" ; applied to all hosts

[host "one"]
  addr = "10.0.20.10:22"
  cmd = "./notify.sh"
  cmd = "./cleanup.sh"

[host "two"]
  username = "admin"
  identity = "~/.ssh/admin_rsa"
  addr = "10.0.20.11:22"

[deploy "get-hostname"]
  host = one
  host = two
  cmd = hostname

[build "initialize"]
  cmd = ./init.sh
  cmd = ./update.sh
  cmd = echo "initialized"

Usage

Usage of ./bin/hap:
--dry=false: Show commands without running them.
-f, --file="Hapfile": Location of a Hapfile.
--force=false: Force build even if it happened before.
--help=false: Show help
-h, --host="": Host to use for commands. Use glob patterns to match multiple hosts. Use --host=* for all hosts.
-v, --verbose=false: [deprecated] Verbose mode is always on

Available Commands:
hap build	        Run the builds and commands from the Hapfile.
hap c <command>		Run an arbitrary command on the remote host.
hap create <name>	Create a new Hapfile at <name>.
hap deploy <name>	Run the named deploy defined in the Hapfile.
hap exec <script>	Execute a script on the remote host.
hap push		Push current repo to the remote.

Advanced Usage

Sometimes you want to build more than one host. If the hosts follow a similar pattern you can reference all the hosts with a *. For example, app-01 and app-02 are configured. Then you can build both with hap -h app-* build or hap -h a* build.

Sometimes you have a lot of hosts that you want to manage in clusters. If you create multiple files you can use the --file flag to specify the location of the config. The file can be named anything. For example, hap -f Appfile -h app* push, will push all the app hosts in the Appfile.

License

The BSD License http://opensource.org/licenses/bsd-license.php.