Database Reference
In-Depth Information
Configuring the Client
The client machines will have to be configured to work with the Distributed Replay controller. Begin by installing your
clients to the different machines. For illustration purposes only, I'm running everything on a single machine; however,
the setup is no different if you use multiple machines. You need to configure the client to work with a controller, and a
client can work with only one controller at a time. You also need to have space on the system for two items. First, you
need a location for working files that are overwritten at each replay. Second, you need room for trace file output from
the client if you want to collect execution data from that client. You also get to decide on the logging level of the client
process. All of this is set in another XML configuration file,
DReplayClient.config
. Here is my configuration:
<Options>
<Controller>DOJO</Controller>
<WorkingDirectory>C:\DRClientWork\</WorkingDirectory>
<ResultDirectory>C:\DRClientOutput\</ResultDirectory>
<LoggingLevel>CRITICAL</LoggingLevel>
</Options>
The directories and logging level are clear. I also had to point the client to the server where the Distributed Replay
service is running. No other settings are required for multiple clients to work; you just need to be sure they're going to
the right controller system.
Running the Distributed Tests
So far you have configured everything and captured the data. Next, you need to go back to the command line to run
things from the
Dreplay.exe
executable. Most of the control is accomplished through the configuration files, so there
is little input required in the executable. You invoke the tests using the following command:
Dreplay replay -d c:\data -w DOJO
You need to feed in the location of the output from the preprocessing, which means you need to list the client
machines that are taking part in a comma-delimited list. The output from the execution would look something like
Figure
24-5
.
Figure 24-5.
The output from running DReplay.exe
