Community
Participate
Working Groups
I have seen this on rare occasion, but recently found it to be somewhat reproducible by repeatedly launching the same test. It seems to happen more frequently when using the agent controller as a service, and seems to happen less frequently when using a 1.5 JRE (Sun or IBM), but both of these data points may be anecdotal, and in all cases, it appears that the launch will hang at 60% eventually. In order to reproduce, check out the test-results/test/org.eclipse.tptp.test.testservices.tests project. Start the RAC (either as a service or as a console) Run the test /org.eclipse.tptp.test.testservices.tests/junit/Test.FilesystemServiceTest.testsuite (right click - Run As - Test). When it completes, run it again. Repeat until the launch sequence hangs indefinitely (for me, this happens some time between the 14th and the 30th execution of the test.)
I can consistently reproduce this problem with the new AC using build TPTP-4.2.0-200604271249. The test execution doesn't actually hang but block for a *very* long time (at least a minute). To reproduce the problem: 1) Run the new AC as a *service* pointing the JDK to Sun 1.5 2) Create a manual test suite with a single test case 3) Switch to the behavior page and add an invocation for the test case added 4) Right click on the test suite > Run As > Test Once the manual test client comes up, click on the terminate button and click yes to exit the program. Click on the run toolbar button to launch the test suite again. In my case it blocked for a minute to just bring up the client. The same happens for auto GUI test suite. Navid noticed the same problem with Auto GUI test suite using 0426 driver. I'm moving this up to critical and changing the component to Platform.Communication.
This problem is causing a lot of inconvenience in the automation effort. I can consistently reproduce this problem with AGR on the 2nd run. I'm actually running the new AC on the command line using SUN JDK 1.4. Ali is correct, it seems to block for a really long time without doing anything. I have to restart the new AC every time before running a second test case to get around this problem. I'm going to switch to the old RAC until this problem is fixed. I think this is a serious issue. I'm marking this bug as a blocker because of the problems its causing in the automation effort.
This was being caused by the Agent Controller failing to update its process list to purge executables which it didn't launch but did know about after they had exited. The queryProcessList command was returning PIDs for these processes, but subsequent attempts to get agent details failed. The UI went into a wait state expecting a response that never came. I changed the default polling state to FALSE (matching the RAC), since the compatibility layer doesn't actually poll. Now, when the client queries the process list, the list is scrubbed.
As of TPTP 4.6.0, TPTP is in maintenance mode and focusing on improving quality by resolving relevant enhancements/defects and increasing test coverage through test creation, automation, Build Verification Tests (BVTs), and expanded run-time execution. As part of the TPTP Bugzilla housecleaning process (see http://wiki.eclipse.org/Bugzilla_Housecleaning_Processes), this enhancement/defect is verified/closed by the Project Lead since this enhancement/defect has been resolved and unverified for more than 1 year and considered to be fixed. If this enhancement/defect is still unresolved and reproducible in the latest TPTP release (http://www.eclipse.org/tptp/home/downloads/), please re-open.
