With the system backed up, the content modified as necessary, and the benchmark defined and configured, you can perform the benchmark test.
As WebSphere InterChange Server starts up, it first loads all business object definitions in the repository and then initializes all maps.
WebSphere InterChange Server then starts the coordinators for all defined
benchmarks that are enabled. For each coordinator that starts, an entry
similar to the one below is written to the logging destination (with the name
of the benchmark definition in place of the text
BenchmarkName):
System: Server] [Thread: main (#8316668)] [Mesg: BENCH : BenchmarkName : Waiting for Participants ] [Time: 2001/12/26 09:22:58.102] [System: Server] [Thread: main (#8316668)] [Mesg: BenchMark ************* BenchmarkName : Starting Coordinator ************** ]
WebSphere InterChange Server next initializes all the connectors in the repository, and then initializes all the collaborations in the repository.
WebSphere InterChange Server indicates its readiness by writing an entry
such as the following to the logging destination:
[Time: 2001/12/26 05:40:55.975] [System: Server] [Thread: main (#8316668)] [Type: Info] [MsgID: 20] [Mesg: InterChange Server "CW400DEV" is ready.]
When WebSphere InterChange Server reaches a ready state, any benchmark
participants synchronize with the coordinator of the benchmarks they
participate in. Because collaborations initialize while WebSphere
InterChange Server is starting, any collaborations that participate in
benchmarks synchronize with their coordinator components at this time.
For each participant that synchronizes with a coordinator component, an entry
such as the one below is written to the logging destination (with X
being replaced by the number of participants that have synchronized and
N being the total number of participants):
[Time: 2001/12/26 05:40:55.935] [System: Server] [Thread: Thread-5 (#5138078)] [Mesg: BenchMarkBenchmarkName5: Participant ParticipantName : X out of N has synchronized ]
You should examine the states of the components that participate in benchmarks once WebSphere InterChange Server has been restarted. Follow the steps below to do so.
If any of the components are not active then the load of events generated by the system during the benchmark will fail, resulting in a poor measurement of system throughput and increased cleanup responsibilities, so it is best to check their states first. If any component involved in a benchmark is not active, right-click the component and click the proper menu item to start the component. If a component remains inactive even after an attempt to start it, there may be issues with the content that need to be resolved.
Once you have made sure that all of the components involved in the benchmark are active, you should start the client participants.
To start clients that are not being emulated you should take the normal steps, such as using the connector_manager.sh script in a Solaris environment, using a shortcut or service in a Windows environment, or using the Object Activation Daemon. To start clients that are being emulated you should use the appropriate client emulator created in the section "Creating client emulator startup scripts".
When you start up client participants they synchronize with the coordinator
component. For each client participant that synchronizes with a
coordinator component, an entry such as the one below is written to the
logging destination (with X being replaced by the number of
participants that have synchronized and N being the total number of
participants).
[Time: 2001/12/26 05:40:55.935] [System: Server] [Thread: Thread-5 (#5138078)] [Mesg: BenchMarkBenchmarkName5: Participant CollaborationObjectName : X out of N has synchronized ]
After all the required participants have synchronized an entry such as the
one below is written to the logging destination, indicating that the benchmark
has begun:
[Time: 2001/12/26 08:43:02.042] [System: Server] [Thread: VBJ ThreadPool Worker (#6231419)] [Mesg: BenchMark : BenchmarkName *********** START BENCHMARK ************]
Benchmarks execute for as long as they are configured to. During the execution you may use the System Monitor or view the logging destination to observe its progress. When the benchmark finishes it automatically shuts down the client participants.
After a benchmark finishes executing the results are output to either the location specified in the benchmark definition, or to the logging destination of WebSphere InterChange Server if no location is configured. If the results were output to a dedicated file then you can view them by opening the file or by following the instructions in the section titled "Viewing benchmark results". If the results were output to the logging destination of WebSphere InterChange Server then you must view them in that location.