Compose a 3-tier Application (UI)

Adding the database tier (MySQL)

Add a node for the database by searching the palette for the type named io.cloudsoft.smart-mysql-node . You can then drag & drop it onto the application node. Alternatively, you can click on the entity in the palette and click the button Add to application.

Once the entity is added into the graph, a new node will appear and the spec editor will then be displayed. Set a name by entering My DB (MySQL) instead of io.cloudsoft.smart-mysql-node. Do the same for the ID by entering db instead of the generated ID.

Finally, edit the mysql_creation_script configuration key and set it to the following value:

classpath://io.cloudsoft.brooklyn.tosca13.tosca-3-tier:resources/creation-script-mysql.sql

Your screen should now look like something similar to this:

Note: io.cloudsoft.smart-mysql-node attempts to download MySQL from the public internet and install it if it is not pre-installed on the VM. Also, this type is suitable for deployment on an Ubuntu 20 VM.

Adding the Web Tier (Cluster of Tomcat)

Add a node for the web cluster by searching the palette for the entity named DynamicCluster or the full type name, i.e. tosca.entity.DynamicCluster and drag & drop it onto the application node.

Once the entity is added into the graph, a new node will appear and the spec editor will then be displayed. Set a name by entering My Cluster instead of Dynamic Cluster. Do the same for the ID by entering cluster instead of the default value.

On the Configuration accordion, set the suggested Initial size (cluster.initial.size) configuration key to: 2.

Set the Member spec (dynamiccluster.memberspec) configuration key by clicking on No spec set. The palette will once again be displayed. Select io.cloudsoft.smart-tomcat9-node and enter the following configuration keys (Configuration accordion):

• tomcat_deployment with value

  classpath://io.cloudsoft.brooklyn.tosca13.tosca-3-tier:resources/tosca-hello-world-sql-webapp.war
  

• tomcat_db_url with value

  $brooklyn:component("db").attributeWhenReady("db_instance_url")
  

• tomcat_db_driver with value

  $brooklyn:component("db").attributeWhenReady("db_instance_driver")
  

Optionally you can rename the spec node from io.cloudsoft.smart-tomcat9-node to Apache Tomcat Member. It is recommended to do so, for more visibility in the App Inspector.

Note: io.cloudsoft.smart-tomcat9-node attempts to download Tomcat 9 from the public internet and install it if it is not pre-installed on the VM. Also, this type is suitable for deployment on an Ubuntu 20 VM.

The spec editor should now display something similar to this:

Go back to the web cluster entity by clicking on the My Cluster cluster node on the graph. Head over the Enrichers accordion and click on the “+” icon. Similar to the cluster member spec in the previous step, search and select the Aggregator enricher and clearing any filters if necessary, add the following configuration keys (Configuration accordion):

• enricher.sourceSensor with value

  $brooklyn:sensor("app.endpoint")
  

• enricher.targetSensor with value

  $brooklyn:sensor("addr.list")
  

• enricher.aggregating.fromMembers with value

  true
  

Head over the Enrichers accordion again and click on the “+” icon and add Joiner with the following configuration keys (Configuration accordion):

• enricher.joiner.minimum with value

  $brooklyn:config("cluster.initial.size")
  

• enricher.sourceSensor with value

  $brooklyn:sensor("addr.list")
  

• enricher.targetSensor with value

  $brooklyn:sensor("addr.list.string")
  

• enricher.joiner.separator with value

  ,
  

The canvas and the spec viewer should now display something similar to this:

Add a node for the web cluster by searching the palette for the entity named Load Balancer (nginx) or the full type name, i.e. tosca-nginx-node and drag & drop it onto the application node. Set a name by entering My Load Balancer (nginx) instead of Load Balancer (nginx) for the tosca-nginx-node. Modify the ID by entering nginx-instance instead of tosca-nginx-node.

For the nginx-instance node, on the Configuration accordion, search for:

• nginx.sticky configuration key and set the switch to false
• addr.list configuration key and enter the following value:

  $brooklyn:entity("cluster").attributeWhenReady("addr.list.string")
  

Note: tosca-nginx-node attempts to download NgiX from the public internet and install it if it is not pre-installed on the VM. Also, this type is suitable for deployment on an Ubuntu 20 VM.

Add a tosca.nodes.Compute to the application and set its id to nginx-host. Make this node a host for nginx-instance. Click on thenginx-instance node and in the spec editor expands the Requirements tab and set the host to nginx-host.

To do this, click on the nginx-host node and in the spec editor, edit the capabilities property. Set it to the following value:

{
  "cloudsoft": {
    "properties": {
      "required.ports": "$brooklyn:component(\"nginx-instance\").config(\"app.port\")"
    }
  }
}

The canvas and the spec viewer should now display something similar to this:

Setting the Application Title, Location and Propagate Valuable Sensor

Click on the application node (where it currently says New application) to bring up the spec editor. Set a name by entering My Web Cluster instead of Unnamed Application.

TOSCA is ordinarily configured with a default location. If this is not done, or you wish to use a different location for this application, open the Location accordion and click on the Attach link. This will display the palette with all available locations you can deploy your blueprint to. Select the one you wish to use.

Finally, head over the Enrichers and click on the “+” icon. Similar to the cluster member spec in the previous step, search and select the Propagator enricher and add the following configuration keys (Configuration accordion):

• enricher.producer with value (or use the DSL editor to target the nginx-instance entity)

  $brooklyn:entity("nginx-instance")
  

• enricher.propagating.inclusions with value

  main.uri
  

The canvas and the spec viewer should display something similar to this:

Note: Although this is a deployable application, for a real life three tier application designed using this model, additional aggregators, propagators and policies might be necessary.

Note: Depending the configuration used for your AMP some extra tabs, type and options might be available.

Adding the database tier (MariaDB)

For the data tier, we will use MariaDB, a popular MySQL-compatible database, which the threetier.zip bundle added to the catalog as cloudsoft-threetier-mariadb. Start by finding this in the palette at left: search for that ID or for the name “Cloudsoft 3T MariaDB”. You can then drag & drop it onto the application node to add it as a child. Alternatively, you can click on the entity in the palette and click the button Add to application.

Once the entity is added into the graph, a new node will appear and the spec editor will then be displayed. Set a name by entering My DB (MariaDB) instead of Cloudsoft 3T MariaDB.

Also give it a nice “reference ID”, a few lines below the title; enter db instead of the random string next to the identity card. This will simplify referring to sensor attributes of this entity when we come to inject the DB address into the mid-tier.

Finally, set the configuration key creation_script_sql_url to the following value:

https://github.com/cloudsoft/hello-world-spring-boot/releases/download/1.0.1/creation-script-mariadb.sql

Your screen should now look like something similar to this:

Adding the web mid-tier: a cluster of Java Spring Boot nodes

For the mid-tier, we will use a “dynamic cluster” which allows resizing the number of nodes created from a “spec”. Search the palette for the entity named DynamicCluster (org.apache.brooklyn.entity.group.DynamicCluster) and drag & drop it onto the application node.

Once the entity is added into the graph, a new node will appear and the spec editor will then be displayed. Set a name by entering My Cluster instead of Dynamic Cluster. Do the same for the ID by entering cluster instead of the default value.

On the Configuration accordion, set the suggested Initial size (cluster.initial.size) configuration key to: 2.

Set the Member spec (dynamiccluster.memberspec) configuration key by clicking on No spec set. The palette will once again be displayed. Select cloudsoft-threetier-java-midtier and enter the following configuration keys (Configuration accordion):

• For war_url, use:

  https://github.com/cloudsoft/hello-world-spring-boot/releases/download/1.0.1/hello-world-spring-boot-1.0.1.jar
  

• For extra_args, open the DSL editor by clicking on the lightning bolt icon. Select the kind Formatted string and then set the following entries:

  • --spring.profiles.active=remote --DB_IP='%s' --DB_USER='%s' --DB_PASS='%s' as the pattern
  • $brooklyn:component("db").attributeWhenReady("host.address") to retrieve the address from our db node, as the first argument (click Add argument then enter, or use the DSL editor again to construct this)
  • brooklyn as the second argument (this is the user name expected in the creation script SQL)
  • $brooklyn:external("brooklyn-demo-sample","hidden-brooklyn-password")) as the password; this references a built-in external secrets store (you can also configure other secrets store) which contains the password specified in the creation script SQL

Then click Done to save. The spec editor should now display something similar to this:

We now need to collect the IP addresses of all nodes created in the dynamic cluster. An “aggregator enricher” will do this. Go back to the web cluster entity by clicking on the My Cluster cluster node on the graph. Head over the Enrichers accordion and click on the “+” icon. Similar to the cluster member spec in the previous step, search and select the Aggregator enricher and, clearing any filters if necessary, add the following configuration keys (Configuration accordion):

• enricher.sourceSensor => $brooklyn:sensor("host.address")
• enricher.targetSensor => $brooklyn:sensor("children_ips")
• enricher.aggregating.fromMembers => true
• enricher.transformation.untyped => list

The canvas and the spec viewer should now display something similar to this:

Adding the load-balancer tier (Nginx)

Add a node for the web cluster by searching the palette for the entity named Cloudsoft 3T Nginx (cloudsoft-threetier-nginx) and drag & drop it onto the application node.

Once the entity is added into the graph, a new node will appear and the spec editor will then be displayed once again. Set a name by entering My Load Balancer (nginx). Do the same for the ID by entering nginx instead of the default value.

On the Configuration accordion, we need to set two values:

• target_ips: $brooklyn:component("cluster").attributeWhenReady("children_ips") (set this as the value, not an entry, by clicking the </> code mode next to “No entries”; this is a DSL expression pointing to the list of IPs maintained by the aggregator enricher we just added at the cluster)
• target_port: 8080 (this is where the Spring Boot app runs)

The canvas and the spec viewer should now display something similar to this:

Setting the application title, location and propagate valuable sensor

Click on the application node (where it currently says New application) to bring up the spec editor. Set a name by entering Cloudsoft 3T Three Tier Application instead of Unnamed Application.

Open the Location accordion and click on the Attach link. This will display the palette with all available locations you can deploy your blueprint to. Select the one you wish to use.

Finally rearrange the children nodes so they go logically from left-to-right as nginx, mid-tier, db.

The canvas and the spec viewer should display something similar to this: