Skip to content

Deploy the platform

Warning

From now on, all the commands for this tutorial will have to be run on the Punch Deployer VM. Use vagrant ssh deployer to connect to this VM.

Configure Environment

Add the punchplatform-deployer.sh script to your PATH:

echo "export PATH=~/punch-deployer-6.4.4-SNAPSHOT/bin:\$PATH" >> ~/.bashrc
source ~/.bashrc

Create deployment folder

Create a folder to add deployment configuration:

mkdir ~/pp-deployment-conf
cp ~/punch-deployer-6.4.4-SNAPSHOT/examples/platforms/getting_started_deployer/deployment-conf/punchplatform-deployment.settings ~/pp-deployment-conf
cp ~/punch-deployer-6.4.4-SNAPSHOT/examples/platforms/getting_started_deployer/deployment-conf/resolv.yaml ~/pp-deployment-conf

The punchplatform-deployment.settings file describes your target platform. It specifies what, where and how components should be installed on your target nodes.

The resolv.yaml files adds parameters to your configuration files. In this tutorial, we will use it to make standalone configuration works on a 3-node platform.

Finally, set the mandatory PUNCHPLATFORM_CONF_DIR variable to point to your deployment configuration folder:

echo "export PUNCHPLATFORM_CONF_DIR=~/pp-deployment-conf" >> ~/.bashrc
source ~/.bashrc

Generate inventory

Execute:

punchplatform-deployer.sh --generate-inventory
What does is do ? It reads your punchplatform-deployment.settings in the $PUNCHPLATFORM_CONF_DIR and generates a complete set of Ansible inventories to deploy the corresponding components on your target servers.

You can check the newly generated inventories in $PUNCHPLATFORM_CONF_DIR/generated_inventory. The main inventory is current-deployment.inv. The other inventories are specific for each component.

Deploy Zookeeper

To deploy Zookeeper use:

punchplatform-deployer.sh --deploy --tags zookeeper

Zookeeper will be deployed on your target servers according to the configuration specified in your punchplatform-deployment.settings:

cat $PUNCHPLATFORM_CONF_DIR/punchplatform-deployment.settings | jq .zookeeper

It returns the part of the file where Zookeeper is configured.

{
  "zookeeper_version": "apache-zookeeper-3.7.0-bin",
  "zookeeper_nodes_production_interface": "eth1",
  "zookeeper_childopts": "-server -Xmx128m -Xms128m",
  "clusters": {
    "common": {
      "hosts": [
        "server3"
      ],
      "cluster_port": 2181,
      "punchplatform_root_node": "/punchplatform-primary"
    }
  }
}
Here we can see that Zookeeper is deployed on server3, with a cluster port opened on 2181.

Deploy other COTS

You can use --tags to deploy components one by one. You can also deploy all the components without using this option.

Let's deploy the other components required for the rest of this tutorial.

punchplatform-deployer.sh --deploy -t zookeeper,kafka,shiva,operator,elasticsearch,kibana -e @/home/vagrant/pp-deployment-conf/deployment_secrets.json

Info

The -e @/home/vagrant/pp-deployment-conf/deployment_secrets.json specifies a file containing credentials. In this tutorial, the file only has the credentials to deploy Elasticsearch with admin:admin. On a production platform, it contains all the credentials and is encrypted with Ansible Vault.