Redwood Platform Application Server
Redwood Platform is a web application server distributed by Redwood that includes the Redwood Server war file ready to be deployed. The application server is based on Apache Tomcat with a number of modifications, Apache Tomcat is not supported without these modifications. It is by far the most tested application server for Redwood Server and as such the recommended platform.
JAR Libraries
The JAR libraries shipped with Redwood Platform in the <install_dir>/j2ee/cluster/global/tomcat/lib are not to be modified, no new JAR file is allowed to be added with the exception of the SAP JCo library as Redwood is not allowed to distribute that. Any JAR file that is not recognized by the application server will be removed without warning from that directory, for stability reasons. You are not to alter JAR files in that directory without specific instructions from Redwood support services; please follow the instructions provided carefully.
JDBC Drivers
The JDBC drivers shipped with Redwood Platform have been extensively tested and validated against the supported databases. If for any reason you wish to update a JDBC driver, you import the JDBC driver into a library and use that to connect to the database.
Configuration
Since Redwood Platform is a modified Apache Tomcat application server, most Apache Tomcat documentation topics also apply to Redwood Platform. Redwood recommends you use the present documentation to configure the application server to your liking; if instructions for a specific configuration cannot be found in this documentation, please consult Redwood support services who will be happy to help.
Redwood Server Web Application
Unlike most JEE web applications, Redwood Server is composed of both a front-end and a back-end; the application is not merely a user interface but also the software that handles all the back-end load. The back-end usually uses most resources to connect to remote systems, compute start times and scheduling the workload across your data center. This means that unlike traditional web application servers, most resources are used by the back-end and not the user interface. Redwood Platform has built-in support for clustering, however, since the primary node will be performing most of the background tasks, clustering does not necessarily confer greater performance; there is communication overhead between the primary node and the secondary nodes in a cluster. Active-active cluster scenarios are not recommended; Redwood Server benefits from more CPU cores and RAM. Multiple cluster nodes on a same host duplicate RAM and CPU usage on top of the communication overhead and are thus not recommended.
Active-passive cluster scenarios are recommended for high availability; they come at no extra licensing costs provided that both nodes have the same hostname and port.
See Also
onsiteTopic