You will want to know what Engine Room data means for a load test using Rational Performance Tester (RPT) and here I will explain how to capture the engine room data for data analysis or troubleshooting RPT agent related issues:
RPT Engine Room data
You can monitor the "RPT Engine Room" periodically on the agent locations to make sure that no unusual conditions are occurring in the location.
The Engine Room data is most useful if you have several samples as the workload increases. Like capture the data in regular intervals , for example for a 500 user load, you can capture the data for 100, 150 , 200, 250 users in HTML format. You can use the RPT agent "Engine Room" to try and deduce what is happening in cases where agents get struck or hung state or fails with driver failed error or when the majordomo service get terminated unexpectedly.
1.Collect the Engine Room data as follows:
You can reach the Engine Room from the RPT workbench by accessing the URL "http://<agentname>:<port>/" Where ,
- 'agentname' is the either the hostname or the IP address of the RPT agent machine where the load generator agent is installed and running &
- 'port' is the value the RPT agent uses to communicate with RPT workbench.
Your RPT schedule can have one or more agent location's. The first location started on a computer will use port 1903; the subsequent locations will use dynamically generated port numbers which you can find in the rptport.dat file in each location's deployment directory. (the deployment directory can be obtained from the Performance Schedule>Schedule element details from the RPT workbench from the controller machine. In the below screenshot, you can see the deployment root directory for the respective agent location is C:\temp
2.How to capture :
Once the schedule starts, bring up a browser on both/all of the agents systems and point to "http://<agentname>:<port>/" you can get information about the current number of users. It also shows the state of engine threads -- they should be OK, WORKING or IDLE.
You can save the data as an HTML file by using the browser's File->Save As option.
4.Interpretation of the collected data :
The data is displayed in four sections
- Engine Counters (Various counters such as CPU, JVM Heap, I/O?etc)
- Subsystems (need info)
- Runner (need info)
- Actions (need info)
Users are able to collect information about the current number of user?s and the state of engine threads (The state should be in the OK, WORKING or IDLE state). Deadlocked threads can be a the underlying problem for agents stopping in mid run. The current state of all active actions are also displayed.For larger
runs, users can focus on one of the user's and track any type of unexpected behavior.
Apart from the engine room log (*.html file) you can always zip the 'deployment_root' directory from the agent machine and share it with IBM Support team for any troubleshooting discussions via a PMR. This will help with analysis data for what happens during a agent run.
The alphanumeric directory is created for each run and will have different timestamps. Note the timestamp at the time of execution to locate the exact execution data directory. You will find numerous file types listed from the execution data.