Understanding Jolt Internet Relay
This section discusses:
Jolt Internet Relay architecture.
A Jolt Internet Relay example.
Jolt Internet Relay Architecture
Image: Jolt Internet Relay Architecture showing the Jolt Relay sending messages through a firewall to the Jolt Relay Adapter
Jolt Internet Relay provides an environment in which the PeopleSoft web server and application server can be further decoupled. This provides greater security at sites where security is of utmost importance. Jolt Internet Relay routes messages from a Jolt client to a Jolt Server Listener (JSL) or Jolt Server Handler (JSH), and eliminates the need for the JSL, JSH, and Tuxedo application to run on the same machine as the web server. Communication takes place between the JRLY and JRAD elements rather than between the Jolt client and JSL/JSH processes. Traditionally an application server domain opens between 2 and 6 ports for such communications. The use of Jolt relay restricts this to one port per domain on the application server machine. This enables an administrator to open just one port on the application server machine. The following diagram illustrates this feature:
Jolt Internet Relay consists of two elements: Jolt Relay (JRLY) and Jolt Relay Adapter (JRAD). It's important to understand the difference between these two elements.
JRLY consists of a standalone program and configuration file; the program runs on the same machine as the web server. JRLY receives Jolt messages from a PeopleSoft web application and routes those messages to JRAD on the application server. It receives the Jolt message through one port, the LISTEN port, and connects to the JRAD by using another port, the CONNECT port. JRLY is sometimes referred to as a front-end relay.
JRAD runs on the same machine as the application server. It's configured on the application server domain as part of the PeopleSoft PSADMIN domain configuration procedure. JRAD listens for JRLY messages on its LISTEN port and transfers the message to the JSL or JSH. JRAD is sometimes referred to as a back-end relay.
Note: Implementing Jolt Relay can impede performance. Always perform testing with typical production system load to ensure it will meet your service level requirements.
Keep the following points in mind as you configure the Jolt Internet Relay components:
The jrly binary and its corresponding jrly.config file must exist in the same directory. To start multiple Jolt Relays on a machine, copy the jrly binary and jrly.config into each subdirectory, modify the parameters in the jrly.config file, and start Jolt Relay. On Windows, you can define multiple Jolt Relay services on a machine.
You can start the JRLY process before or after you start JRAD. The JRLY attempts to connect to JRAD on the client request. If the JRLY is unable to connect to the JRAD, the client is denied access and disconnected. The connection will be retried upon the first use of PeopleTools..
If you're installing Jolt Internet Relay on UNIX and anticipate a large number of concurrent connected clients, increase the file descriptors limit before running the JRLY executable.
At runtime, if you get the following message:
[FRi JUNl 6 20:25:11 1997] JRLY:accept():accept failed, err no: 23, strerror: File table overflow
PeopleSoft recommends that you increase the MaxUSERS kernel parameter and regenerate the kernel.
If you're unable to connect, make sure that you check the following items:
Port numbers do not match.
Print out the jrly.cfg file and the psappsrv.cfg file and compare the port numbers that you specified.
Make sure that the application server is running.
Make sure that JRLY is running.
Jolt Internet Relay can be installed on an intermediate machine rather than the web server machine if necessary. This extra level of indirection can cause peformance degradation.
Make sure that JRAD is running on the application server and that you configure JRAD using PSADMIN.