How to configure a new service

• In order to create a service template called fancyPackage (Without going into much details) you've to do:
  • Publish service binary data
    • ofcourse only if you have binary data to use which you want to repackage yourselves
    • Initiate a local git repo anywhere in your file system called fancyPackage.
    • Add the binary data files to this repo
    • Push this repo to e.g. Remote Binary Repo/Aydo
  • Create service templates
    • create a directory under your ays metadata repo e.g. /opt/code/github/jumpscale7/ays_jumpscale7/ called fancyPackage
    • Inside the service template directory, add 3 files (actions.py, service.hrd, instance.hrd)
    • Those metadata files draw the life cycle plan of the service from creation to monitoring.
    • Metadata will contain the URL for binary data repo for that service on Remote Binary Repo/Aydo

In the next sections we'll explain in details how to configure and create a new service.

A service has 2 HRD files and up to 3 python files. HRD files are where configuration is kept, python file describe the actions possible of the services.

• service.hrd : instance specific configuration
• instance.hrd: common configuration for all instance of this service
• actions_tmpl.py : contains the actions possible for a templates directly without the need to create an instance of the service to call them. Can be seen as "static" method of a service.
• actions_mgmt.py : contains actions executed from the management location.

• actions_node.py : contains actions executed remotely on a node.

  • contains Actions class with predefined functions that do certain actions
  • here're the steps involved in a service installation
    • At management location
      • init : instantiate service, check dependencies and install them if possible
      • prepare
      • configure
      • check_up : Checks if process already running (from previous installation)
        • If process running try execute stop for graceful stop
        • execute check_up to check if process still running
        • If process still running try execute halt for hard stop
    • start using the config files to start the application
    • check_up to check process is started
    • monitor to check local application is running & healthy
  • If the service is installed on a remote node, after management actions. method from action_node.py are executed on the remote node.
    • configure
      • check_up : Checks if process already running (from previous installation)
        • If process running try execute stop for graceful stop
        • execute check_up to check if process still running
        • If process still running try execute halt for hard stop
    • start using the config files to start the application
    • check_up to check process is started
    • monitor to check local application is running & healthy

You can read more about the HRD format here.