Overview

Repository for automated deployment/configuration of computing nodes associated with networkplanner.

To deploy to a network planner node use:

# setup baseline packages on machine
ssh <host> 'bash -s' < bootstrap.sh
# setup network planner specific packages, create np user and directories
fab -H <host> provision:system_type=<system>,branch=<np_devops_branch>
# deploy network planner as the np user (note that host_np_user should login as np user)
fab -H <host_np_user> deploy:system_type=<system>,branch=<np_branch>,config_env=<my_config_env.yaml>

Where <system> is one of:

• ss - Single Server Mode
• cs - Cluster Server that "manages" clusters
• cp - A cluster processor computing node

For cluster processors, you can pass multiple comma delim'd hosts to fab and the -P option to provision or deploy in parallel.

Also includes automated backup utils.

Main Components

• fabfile.py - fabfile which orchestrates deployment
• chef-repo - the chef configuration for setting up nodes
• <server-type>.json - files defining chef components of each server type
• backup.sh - script defining backup process

Caveat Emptor

• You should have a reasonable understanding of Network Planner and it's architecture before using this (see: github.com/SEL-Columbia/networkplanner)

• Running the deploy task requires a yaml file (see sample.yaml) that contains important configuration information. See fabfile.py for default file names. It can be overridden by a config_env command line parameter.

• See fabfile.py for details