Deploying a Model

Deployment is the process of transferring a model and its associated files and settings to the production server or servers where it's going to run. You deploy a model from the Design Tool or you can use command-driven deployment.

After a models is deployed, it is visible in the Models view of the Administrative Console. To open the Models view, from the View menu, select Models.

You can use the Design Tool to deploy a model with one configuration to one server, in this case a development server, running on the same computer as the Design Tool,for test purposes. In a command-driven deployment, you use Verastream Host Integrator's deployment commands from a command line or in a batch or shell file to deploy a model with one or more configurations to one or more production servers.

If your integration solution includes a Web application created with Verastream's Web Builder, you must also separately deploy the Web application files. See Deploy Web Applications in the Web Builder help for information.

Using the Design Tool to deploy

If you are deploying a model with one configuration to the development server for test purposes, it's easiest to use the Design Tool.

With this approach, you use the Design Tool's Deployment Options command on the File menu to designate any session pool and model variable list settings. Next, you use the File menu's Deploy to Local Server or Deploy to Remote Server command to deploy the model. If Verastream security is enabled on the server, you are prompted for credentials. Unlike the command-driven method described below, you can only deploy the model to one server, using one model configuration. See Using the Design Tool to Deploy a Model in the Design Tool help for more information.

Using commands to deploy

If you are deploying a model with one or more configurations to one or more production servers, you should use Verastream Host Integrator's deployment commands from a command line or in a shell or batch or shell file.

With this approach, you create a model package and then use commands or a batch or shell file to deploy that model package. Your batch file can automate the deployment to multiple Host Integrator Servers. In the model package, the model file can be combined with event handler .JAR files and descriptors that tell the servers how to provide access to the model -- such as via a session pool of a certain size, or by using specific requests for a new session. See About Model Packages and Using Commands to Deploy a Model Package for more information about this deployment method.

For more information

There are topics on deploying model packages and working with descriptor files and commands in the Design Tool help. See Deploying Model Packages.

Updating models deployed with earlier versions of Host Integrator

If you're upgrading from an earlier version of Host Integrator, you may have a lot of model or session pool configuration information in the Administrative Console. If you want to update a model without affecting existing configuration information, do not include descriptor files in your model package.

Once you deactivate a model that has configuration information in the console, that configuration information is lost. So if you're creating new models, you should put such information in a configuration descriptor file and then build it into your model package.

 

Related Topics
Bullet How to Use Host Integrator