aws-eapamq-demo

This repository contains reference configurations that can create the needed resources on AWS and deploy on them a cross-region clustered Jboss EAP service with the JMS remoting to a cross-region AMQ broker cluster. The demo relies on ansible middleware automation collections for wildfly/EAP and activemq/AMQ Broker.

The final architecture looks like:

Prerequisites

• Red Hat Enterprise Linux (RHEL) 8 controller host (or Fedora works too); preferably a bastion ec2 instance
• Red Hat Ansible >= 2.11 / python >= 3.9
• python dependencies installed on the controller host pip install -r requirements.txt
• AWS-CLI installed on the controller host
• an AWS account capable of:
  • vpc:*
  • ec2:*
  • route53:*
  • efs:* (only if setting up backup amq brokers with shared filesystem)
• an ssh keypair installed in AWS to use for creating EC2 instances
• if exposing the loadbalancer publicly, the TLS certificate and key in the files/ directory, named public_fqdn.key and public_fqdn.crt

Install Ansible Dependencies

ansible-galaxy collection install -r requirements.yml

Running on AWS cloud

Infrastracture

1. Make sure ansible and AWS account are operational:

ansible-inventory --list

This should return empty before running the infra.

2. Execute the main play:

ansible-playbook infra.yml

Note: open ansible.cfg config file and edit remote_user and remote_privkey accordingly to your configuration.

3. Now inspect the inventory:

$ ansible-inventory --list

    "all": {
        "children": [
            "amq",
            "aws_ec2",
            "eap",
            "loadbalancer",
            "site1",
            "site2",
            "ungrouped",
            "us_east_1",
            "us_west_2"
        ]
    },
    "amq": {
        "hosts": [
            "site1-amq1",
            "site1-amq2",
            "site2-amq1",
            "site2-amq2"
        ]
    },
    "eap": {
        "hosts": [
            "site1-eap1",
            "site1-eap2",
            "site2-eap1",
            "site2-eap2"
        ]
    },
    "loadbalancer": {
        "hosts": [
            "site1-loadbalancer",
            "site2-loadbalancer"
        ]
    },
    "site1": {
        "hosts": [
            "site1-amq1",
            "site1-amq2",
            "site1-eap1",
            "site1-eap2",
            "site1-loadbalancer"
        ]
    },
    "site2": {
        "hosts": [
            "site2-amq1",
            "site2-amq2",
            "site2-eap1",
            "site2-eap2",
            "site2-loadbalancer"
        ]
    },

[...]

We use a dynamic ansible inventory that relies on AWS tagging to build the node list and determine configurations, available at inventory/myaws_ec2.yml; and a default ansible configuration at: ./ansible.cfg; by default, ansible ssh connectivity uses EC2 instances public IP addresses, but, if you intend to run on a separate EC2 bastion (recommended), we advice the edit the dynamic inventory like so:

[...]
compose:
  ansible_host: private_ip_address

in order to use instance's private IP address. This saves time, bandwidth and money.

Service deployment

1. Create a var-file containing your RHN credentials:

$ cat rhn-creds.yml
rhn_username: '<username>'
rhn_password: '<password>'

2. Execute the full deployment play

ansible-playbook -e @rhn-creds.yml deploy.yml

Verify deployed demo

1. Open your browser to: https://<your_loadbalacer>/iot-example/ for the consumer UI
2. curl https://<your_loadbalacer>/iot-example/DeviceServletClient?msg=500&interval=50 to produce 500 messages with 50 msec delay
3. watch the consumer UI populate with data
4. (optional) check metrics grafana

License

Apache License v2.0 or later

See LICENCE to view the full text.

Authors

• Guido Grazioli ggraziol@redhat.com
• Shaaf Syed sshaaf@redhat.com
• and the rest of Middleware Automation team