Default Report from RPT not sent to RQM.
I have a test case linked to a test schedule in RQM. I ran the TC from RQM, and while the test was running I saw the summary report stats being populated in RQM. When the test ended, in Execution results there is no report i.e.
The verdict is inconclusive because the performance schedule did not have any performance requirements evaluated during the last user stage. Please examine the attached report and set the verdict appropriately. Consider updating the performance schedule to include performance requirements so that the verdict can be set automatically for you based on whether all performance requirements pass or not. No reports available. Any ideas why this is not in RQM, but I do see the report in RPT. |
15 answers
What version of RPT and what version of RQM?
This problem could be on the client side (ie. RPT adapter) or on the RQM server side. There is logging within RPT that should help identify if a problem occurred in the RPT adapter. In RPT go to Windows->Show View->Error log. On the toolbar of the error log click the red X to clear the log. Then run a RPT script via RQM, allow it to complete and see if any meaningful errors show in the Error Log view. Unfortunately, the logging could use some improvements to debug this specific scenario. So hopefully we'll see some error message in the Error log view that helps understand the issue. If there isn't any traces of an error on the RPT adapter side, we could use the RQM URL utility to access the execution results in the RQM server. The URL tool is documented here: https://jazz.net/wiki/bin/view/Main/RQMURLUtility And is available for download on the "All Downloads" page of your RQM version on jazz.net. The URL that you would pass to this tool would be like: https://localhost:9443/jazz/secure/service/com.ibm.rqm.integration.service.IIntegrationService/resources/Quality+Manager/executionresult/urn:com.ibm.rqm:executionresult:1 where '1' is the ID that you see in the RQM UI when listing execution results. The RQMURLUtility will produce an .xml file for the execution result and that file would contain "attachment" elements if RPT actually did attach a result. If there aren't any attachment elements, then we know something went wrong on the RPT side. |
|
Kurt,
Thanks very much for your response I really appreciate it. Just to bring you up to speed, we had our perofmance guy implement the performance requirements in RPT, so now we are getting a pass message when those requirements are met after a test, but there is still no attachment in the result details in RQM; The one question we have is are there certain permissions that the RQM adapter ID needs to have in order to 1.Update the TER, and 2. be able to attach the default report in RQM. On a side note another issue we are having is that when we run the adapter as a service there are no performance counter information gathered (the ones we selected), but then the adapters is started through the workbench they are gathered, any ideas on this one? (in both instances tests are initiated from RQM). We prefer a service in order to enable scheduling during off hours via RQM. RPT version: 8.1 Rational Quality Manager; version 2.0.1 iFix 1 build: I20100426_1041 Thanks again for your response. Regards, Homer What version of RPT and what version of RQM? |
|
Also, we did look at the log you specify(through RPT), there is an exception being thrown in method: beginTestLogTransfer and it says there is no stack trace availabe. Here are the last few lines of the log:
!ENTRY com.ibm.rational.test.common.schedule.execution 1 0 2010-06-16 11:25:29.243 !MESSAGE RPTA0000W SmartLoad statistics for SSPM_001c !ENTRY com.ibm.rational.test.common.schedule.execution 1 0 2010-06-16 11:25:29.243 !MESSAGE RPTA0000W SmartLoadManager did not complete all stages, stop requested true reason All Agents Inactive, interrupted false !ENTRY com.ibm.rational.test.common.schedule.execution 1 0 2010-06-16 11:25:29.243 !MESSAGE RPTA0000W Stage 1: 1 users, duration 600 seconds, actual indeterminate number of users, duration 550 seconds, Stop requested reason All Agents Inactive !ENTRY com.ibm.rational.test.common.schedule.execution 1 0 2010-06-16 11:25:29.243 !MESSAGE RPTA0000W SmartLoadManager exiting normally !ENTRY com.ibm.rational.test.common.schedule.execution 2 0 2010-06-16 11:26:16.259 !MESSAGE RPTA0000W ScheduleExecutorStub: beginTestLogTransfer() Kurt, What version of RPT and what version of RQM? |
|
Those messages about the SmartLoadManager did not complete all stages... are normal. I also considered your idea about permissions. The execution result is being created, so the user seems to have the appropriate permission for that. I also looked for permissions relating to attachments and couldn't find any.
Here's the next thing to do: 1) Go to Windows->Preferences->Logging. 2) On the "Loggers" tab, look for com.ibm.rational.test.lt.execution.export and com.ibm.rational.test.lt.rqm.adapter. Set the log level of these 2 components to ALL. 3) Open the Error log view, click the red X on the toolbar of Error Log view. 4) Use RQM to run the schedule and reproduce the issue. 5) Use the export log button on the toolbar of the Error Log view to save the RPT log file. 6) Email me that log file. I assume when you run the RPT schedule from RQM, you see the default report open in the RPT UI and when it ends that report automatically closes. That's how it normally works. |
|
Hi Kurt,
We decided to try a few things and we belive we have identified what was causing the issue. Our Project areas was one that contained spaces so when configuring the adapter in the Workbench we replaced the spaces with "+" signs. Funny enough, when we started the adapter it connected and worked flawlessly even sending summary report numbers back to RQM during the run, but never attaching reports at the end of the run. Well after removing the "+" signs and re running, we were able to get reports attahced in RQM. Luckily we had looked more closely at the RQM logs and saw invalid project area messages that corresponded to times we were running Test cases from RQM. The only outstanding issue we have is when running the adapter as a service performance counters are not captured. any idea on this one? Thanks again for your help. Those messages about the SmartLoadManager did not complete all stages... are normal. I also considered your idea about permissions. The execution result is being created, so the user seems to have the appropriate permission for that. I also looked for permissions relating to attachments and couldn't find any. |
|
Glad to hear you got past one problem. Just to warn you there is a similar issue you will see if you ever rename a project area. Here's a related technote:
http://www-01.ibm.com/support/docview.wss?uid=swg21394276 The issue was fixed in RPT 8.1.1. In reference to your other issue, you say that when running as a service "performance counters are not captured". What does that mean? Are you talking about the run status counters which are shown in the RQM web UI when the script is running? Does the adapter connect when running as a service and show up in the RQM adapter console as "available"? When running as a service and you execute a script, does the status in RQM ever show that the run completed and the RPT report is available from the RQM web UI? I believe you're talking about the performance counters on the status page while the script is running. If that's the case, I believe the adapter isn't really available to RQM and the adapter isn't really executing the script. When running as a service you want to view the file <RPT_INSTALL_DIR>\RPT-RST_RQMAdapter\logs\adapter.log to better understand what the adapter service is doing. It will tell you things such as "adapter successfully connected", "adapter executing script XYZ...". If the adapter isn't really available to RQM, then the performance counters on the run status page of RQM web UI will not show up, as you described. It's worth mentioning there is a delay between the time you start the service and the time the service is available to RQM. The service isn't fully started until you see a "successfully connected" message in adapter.log. |
|
Thanks for the heads up on the new issue.
When speaking about performance counters we are targetting the one on the RPT machine which captures counters from the targeted server (credentials saved at the location level). When the adapter was started through the workbench performance counters for the target machine were captured by RPT. Also: Often, the first test case executed after the RPT-RQM adapter is started (either as a service or through the WorkBench) through RQM will fail . Subsequent execution attempts will work correctly as long as the adapter is not restarted. Thanks for the comments & assistance. Homer Glad to hear you got past one problem. Just to warn you there is a similar issue you will see if you ever rename a project area. Here's a related technote: |
|
Ok, you are talking about "Resource Monitoring" counters. These are the counters that are specified on the "Resource Monitoring" tab of the Performance Schedule and show up on the "Resources" tab of the report.
Fortunately I was able to reproduce the problem you are talking about locally and found an easy solution. The problem can be avoid if you specified actual login credentials for the Windows Service, as opposed to using the default "Local System account". The account that you should use is the same user account that you are logged on as when you tested the adapter from within the RPT workbench. This account should have sufficient privileges to do everything the adapter needs to do. To change the account go to the Windows Service panel, right click the RPT adapter service and say "Properties", then go to the "Log On" tab and choose the "This account" option. The resource monitoring feature of RPT uses Windows Perfmon to collect windows performance counters. Apparently the Local System account doesn't have sufficient privileges to use the performance monitoring services as needed by RPT. Please try this solution out and let me know if it works for you. In the mean time I'll create a defect against RPT to further investigate which privileges are causing this behavior for resource monitoring. I'll also create a Tech Note to inform other customers. |
|
Thanks Kurt,
We will give it a try. Also when configuring the adapter on the RPT machine in the Workbench and specifying the RQM Project workspace, if left blank, the adapter will only show in the default project "Quality Manager" project. We were thinking the functionality of leaving this field blank would make the adapter available to all projects in RQM to enable scheduling for multiple projects on the same RPT machine. But it seems that you have to be specific when designating a single project workspace. Are there any workarounds for this, so we can schedule multiple projects on the same machine with no human intervention? Thanks, Homer Ok, you are talking about "Resource Monitoring" counters. These are the counters that are specified on the "Resource Monitoring" tab of the Performance Schedule and show up on the "Resources" tab of the report. |
|
When you leave the project area blank, that's telling the adapter to use the default project area (i.e. "Quality Manager"). We allow the project area field to be blank because the adapter is suppose to work with older releases of RQM, were project area wasn't even introduced yet (RQM 1.0).
Currently you cannot have 1 RPT adapter work for all RQM project areas at the same time. A single RPT adapter is tied to a specific RQM project area. To service multiple project areas, you need multiple instances of the adapter. This is true for all RQM adapters (Command-line adapter, RPT, RST, RFT, etc...). So the only thing you can do here is have multiple instances of RPT installed, most likely on different machines since the Installation Manager will only allow you to install one instance of RPT on a single machine. Not sure about your hardware resources, but if you have it, consider using a virtual image server. In reference to your other comment about the first test you run after starting the adapter fails. I don't see that problem and would need a more info to help diagnose. What is the error message shown in RQM when it fails? Also after it fails, stop the service and start the RPT workbench and see what's in the Error Log view. I also recommend that you move to RPT 8.1.0.3, there was a bug that would cause RQM initiated runs to never end. It happened intermittently and only for RPT test, as opposed to RPT schedules. That bug was fixed in 8103, so maybe worth you applying the fixpack. |
Your answer
Dashboards and work items are no longer publicly available, so some links may be invalid. We now provide similar information through other means. Learn more here.