Deploying to a running container

Cargo supports deploying to an already running container. This feature is called Hot Deployment). You call it by using the (cargo:deploy) goal (e.g. mvn cargo:deploy).

No need to deploy if you are using Cargo to start your container

If you are using Cargo for starting your container, Cargo will actually be doing a static deployment of all your <deployables> automatically. So, no need to deploy anything manually in that case!

About WAR contexts

Many containers have their specific files for redefining context roots (Tomcat has context.xml, JBoss has jboss-web.xml, etc.). If your WAR has such a file, the server will most probably use the context root defined in that file instead of the one you specify using the Codehaus Cargo deployer.

Using a local deployer

A local deployer is a deployer that deploys deployables on a Local Container (i.e. a container installed on the same machine where the deployer is executing). Thus you'll need to use a local container id in <containerId>. You can check that by reviewing the supported container list and selecting the container you wish to use.

Example of doing a local deploy to an existing configuration:

[...]
<configuration>

  <!-- Container configuration -->
  <container>
    <containerId>resin3x</containerId>
    <home>c:/apps/resin-3.0.9</home> or
    <zipUrlInstaller>
      <url>http://www.caucho.com/download/resin-3.0.9.zip</url>
      <downloadDir>${project.build.directory}/downloads</downloadDir>
      <extractDir>${project.build.directory}/extracts</extractDir>
    </zipUrlInstaller>
  </container>

  <!-- Configuration to use with the container (which will also configure the deployer) -->
  <configuration>
    <type>existing</type>
    <properties>
      [...]
    </properties>
  </configuration>

  <!-- Deployables configuration -->
  <deployables>
    <deployable>
      <groupId>war group id</groupId>
      <artifactId>war artifact id</artifactId>
      <type>war</type>
      <properties>
        <context>optional root context</context>
      </properties>
      <pingURL>optional url to ping to know if deployable is done or not</pingURL>
      <pingTimeout>optional timeout to ping (default 20000 milliseconds)</pingTimeout>
    </deployable>
    <deployable>
      <groupId>ear group id</groupId>
      <artifactId>ear artifact id</artifactId>
      <type>ear</type>
      <pingURL>optional url to ping to know if deployable is done or not</pingURL>
      <pingTimeout>optional timeout to ping (default 20000 milliseconds)</pingTimeout>
    </deployable>
    [...]
  </deployables>
</configuration>
[...]

Using a remote deployer

A remote deployer is a deployer that deploys deployables on a Remote Container (i.e. a container that is running and that has been started externally from Cargo). Thus you'll need to use an id for a remote container in <containerId> and a Runtime Configuration.

Not all containers have a remote Deployer implemented

We haven't finished implementing remote Deployers for all containers yet. Please check if your favorite container has it implemented. If not you'll need to deploy your deployables by defining them in a standalone local configuration as shown in the starting and stopping a container use case or the Cargo Daemon.

Example of doing a remote deploy using a runtime configuration:

[...]
<configuration>

  <!-- Container configuration -->
  <container>
    <containerId>tomcat6x</containerId>
    <type>remote</type>
  </container>

  <!-- Configuration to use with the container (which will also configure the deployer) -->
  <configuration>
    <type>runtime</type>
    <properties>
      <cargo.remote.username>username</cargo.remote.username>
      <cargo.remote.password>password</cargo.remote.password>
    </properties>
  </configuration>

  <!-- Deployables configuration -->
  <deployables>
    <deployable>
      <groupId>war group id</groupId>
      <artifactId>war artifact id</artifactId>
      <type>war</type>
      <properties>
        <context>optional root context</context>
      </properties>
      <pingURL>optional url to ping to know if deployable is done or not</pingURL>
      <pingTimeout>optional timeout to ping (default 20000 milliseconds)</pingTimeout>
    </deployable>
    [...]
  </deployables>

</configuration>
[...]

As you can see the information to connect and do the deployment to the remote container is specified in the runtime configuration (cargo.remote.username and cargo.remote.password). The properties to define are deployer-dependent.

Here's an another example, this time deploying the current project's artifact to a running JBoss 4.2.x container using a default 8080 port and default authorizations:

[...]
<configuration>
  <container>
    <containerId>jboss42x</containerId>
    <type>remote</type>
  </container>
</configuration>
[...]

Just type mvn cargo:deploy. Notice that we haven't specified a <deployer> element nor a <configuration> one: this is because the plugin is smart enough to create default instances for you.

Handling of deployment or undeployment failures

If the deployment, undeployment or redeployment is done without any watchdog, Codehaus Cargo will rely on the error codes return by the container's deployer to detect failures.

This is sometimes not desired: for example, if one calls cargo:undeploy but the target deployable (be it a web application, EAR, etc.) is actually not deployed, then the call to cargo:undeploy will return a failure indicating that the deployble to undeploy is not deployed.

To work around the issue, make sure you define your deployable together with a watchdog. For example:

Deployable with a ping URL
<deployables>
  <deployable>
    <location>${project.build.directory}/${project.build.finalName}.${project.packaging}</location>
    <pingURL>http://localhost:${servlet.port}/mycontext/index.html</pingURL>
  </deployable>
</deployables>

In that case:

  • When you will call cargo:deploy:
    • If the deployer returns a failure, Codehaus Cargo will log a message The deployment has failed, with the error returned by the deployer and a severity INFO.
    • It will then start the watchdog, to check if the deployable is deployed.
    • If the watchdog detects that the deployable is indeed deployed, the overall result is a BUILD SUCCESSFUL
    • If the watchdog detects that the deployable is not deployed, you then get a BUILD FAILURE
  • When you will call cargo:undeploy:
    • If the deployer returns a failure, Codehaus Cargo will log a message The undeployment has failed, with the error returned by the deployer and a severity INFO.
    • It will then start the watchdog, to check if the deployable is not deployed.
    • If the watchdog detects that the deployable is indeed not deployed, the overall result is a BUILD SUCCESSFUL
    • If the watchdog detects that the deployable is still deployed, you then get a BUILD FAILURE

This functionality is available in the Java API, Ant tasks and Maven 3 plugin.