Skip to content

Latest commit

 

History

History
86 lines (74 loc) · 5.3 KB

README.md

File metadata and controls

86 lines (74 loc) · 5.3 KB
Raw

Spin up and configure instances for pulp server via ansible

This repository contains multiple possible configurations for setting up pulp on ec2 including running pulp automation and jenkins deployment.

Prerequisites

  • ec2 credentials and RHN credentials.
  • ec2 keypair in selected region.
  • ansible 1.9.x installed on your computer

Setup single pulp node on ec2, without running automation

  • Clone repository using git clone: git clone https://github.com/peterlacko/pulp-automation-ci
  • Copy your private key to pulp-automation-ci/keys directory.
  • Set variables in setEnv.sh file and source the file. Configuration options are described within file itself.
  • Optionally modify general-deploy.yml and general-configure.yml playbooks.
  • To deploy machine in ec2 cloud, run: ansible-playbook -i ec2.py --extra-vars "@global_vars.yml" general-deploy.yml
  • To install and configure pulp on deployed machine run ansible-playbook -i ec2.py --extra-vars "@global_vars.yml" general-configure.yml
  • Now you can access server by it's public dns name, ie. to ssh into it, run: ssh -i <you private key> [email protected]

Playbooks

  • general-deploy.yml -- deploy single ec2 server,
  • general-configure.yml -- configure that server,
  • test-scenarios/ -- playbooks in this directory test various scenarios of pulp deployment, including running pulp-automation
  • automation-{deploy,configure,run}.yml -- run playbooks in test-scenarios directory
  • keypair-gen.yml -- generate new keypair for use in automation, saved to ./keys/
  • security-gropus-create.yml -- create security gropus 'ssh' and 'pulp'
  • automation-runner-configure/deploy.yml -- setup and creates nod for running automation
  • ec2-terminate-all.yml -- serves for terminating specified instances
  • reposerver-{deploy,run} -- deploys and configures machine for building pulp (+plugins) rpms from github and uploads them to repository on s3 bucket

Terminate selected instances

Add instances to terminate to playbook and run:

  • ansible-playbook -i ec2.py ec2-terminate-all.yml -e @global_vars.yml Instances to terminate are specified by it's tag name. To get list of all instancesrun ./ec2.py --list

Config files

  • setEnv.sh -- main deployment configuration file
  • ansible.cfg -- main ansible config file
  • ec2.ini -- ansible dynamic inventory config file
  • global_vars.yml -- reads variables from environment. Usually you won't need to modify the file.

OBSOLETE -Deploy machine, install and configure pulp

  • configure setEnv.sh.template file and source it
  • copy your private key to .
  • Tags: you can specify which tasks from playbook should run by --tags resp which should be skipped by --skip-tags, supported tags for playbooks are in []
  • set up automation runner node as follows
    • ansible-playbook -i ec2.py automation-runner-deploy.yml -e @global_vars.yml
    • ansible-playbook -i ec2.py automation-runner-configure.yml -e @global_vars.yml
  • Comment out test scenarios you don't want to run in automation-{deploy,configure,run}.yml files and then run
    • ansible-playbook -i ec2.py automation-deploy.yml -e @global_vars.yml
    • ansible-playbook -i ec2.py automation-configure.yml -e @global_vars.yml
    • ansible-playbook -i ec2.py automation-run.yml -e @global_vars.yml [run_automation]
  • results will be present in local xml JUnit file
  • for running sigle tests, try your own configuration, run commands, etc, you can still ssh into automation runner or automation node

OBSOLETE - Pulp-automation deployment using ansible

  • each deployment (specified by automation_name, including bucket) should be used for testing of single pulp version (ie. do not mix 2.7-testing with 2.7-devel) where different testing scenarios run from single automation node.
  • ansible is idempotent, thus new nodes won't be deployed if node with given name already exists
  • do not use dash '-' in name of nodes (including automation_name), use underscore instead
  • as testing nodes are currently supported Fedora 20, 22, RedHat 7.1, as automation runner and reposerver Fedora only

Following test scenarios are currently supported:

  • single-node: everything is running on single node. Upon each run pulp and pulp automation is updated to newest version and tests are running. This is simplest deployment, run it first.

OBSOLETE - Usage

  • copy templates of config files from temp/ directory to ansible/ dir
  • export your credentials and set variables in cofig files (see below)
    • export AWS_ACCESS_KEY_ID=<your_access_key>
    • export AWS_SECRET_ACCESS_KEY=<your_secret_access_key>
  • copy your own private key to ./keys/ or generate new keypair using keypair-gen:
    • ansible-playbook keypair-gen.yml -e @global_vars.yml
  • create security groups for simple access and pulp (vars ssh_sg and pulp_sg)
    • ansible-playbook security-groups-create.yml -e @global_vars.yml

OBSOLETE - S3 repositories deploy, config & run

Setup node for building pulp rpms and create & configure S3 bucket as a pulp repo run: