Bugzilla – Full Text Bug Listing
| Summary: | [zoodiscovery][examples] hello example consumer sometimes throws BundleStateException | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [RT] ECF | Reporter: | Scott Lewis <slewis> | ||||
| Component: | ecf.discovery | Assignee: | Wim Jongman <wim.jongman> | ||||
| Status: | RESOLVED FIXED | QA Contact: | |||||
| Severity: | major | ||||||
| Priority: | P3 | CC: | ahmed.aadel, bugs.eclipse.org, wim.jongman | ||||
| Version: | unspecified | ||||||
| Target Milestone: | 3.5.2 | ||||||
| Hardware: | PC | ||||||
| OS: | Windows XP | ||||||
| Whiteboard: | |||||||
| Attachments: |
|
||||||
|
Description
Scott Lewis
Created attachment 197473 [details]
revised DiscoveryActivator class for org.eclipse.ecf.provider.zookeeper provider
This attachment provides a revised implementation of the DiscoveryActivator for zookeeper provider. In the old DiscoveryActivator, a new Thread was created in start(BundleContext) method, and that thread called the DiscoveryActivator.startup method asynchronously (in new thread). I believe that the timing of the execution of this thread could be problematic, and lead to the BundleStatusException reported by this bug (or other exceptions).
The attached DiscoveryActivator uses a ServiceFactory to delay the creation of ZooDiscoveryContainer.getSingleton() method, which loads a number of classes and creates the ZooDiscoveryContainer instance...without using a separate thread to do so asynchronously. This seems safer, as the ServiceFactory registration will complete without BundleStatusExceptions.
I've tested this fix with the zoodiscovery and r-osgi distribution provider. Please review, test and if the fix is acceptable to the zookeeper provider authors, then please commit and release to master. Thanks.
Adding Ahmed and myself. Thanks Scott for the patch. I have updated the code and commited to master. The patch is currently compiling. After this is done I will consume the new zoodiscovery in our own product to test. Added milestone 3.5.2 and added it as target milestone for this bug I have implemented the code but bucky aborts the build because of a time-out. It seems that the discovery tests run for a long period of time [1]. This is different from the last successful build [2] where these tests took a few seconds. Last build completed normally btw but tests take so long that they come close to the alotted build time of 15 minutes. [1] https://build.ecf-project.org/jenkins/job/C-HEAD-discovery.zookeeper.feature/143/console [2] https://build.ecf-project.org/jenkins/job/C-HEAD-discovery.zookeeper.feature/123/console CURRENT TIMES: ------------------------------------------------------------------------ testAddServiceListenerIServiceTypeIDIServiceListenerOSGi... ...OK [40.023s] testAddServiceListenerIServiceTypeIDIServiceListenerOSGiWildcards... ...OK [39.965s] ZooDiscoveryServiceTest.testAddServiceTypeListenerOSGi... ...OK [40.07s] ZooDiscoveryServiceTest.testAddServiceListenerIServiceListener... ...OK [39.953s] etc .. INFO: ...OK [40.016s] INFO: ...OK [39.958s] INFO: ...OK [39.014s] INFO: ...OK [38.966s] INFO: ...OK [39.016s] INFO: ...OK [38.983s] INFO: ...OK [38.966s] INFO: ...OK [39.004s] INFO: ...OK [38.961s] INFO: ...OK [39.007s] INFO: Running test org.eclipse.ecf.tests.provider.zookeeper.ZooDiscoveryServiceTest.testAddServiceListenerIServiceListenerOSGi... Build timed out (after 15 minutes). Marking the build as aborted. FATAL: java.lang.InterruptedException java.lang.RuntimeException: java.lang.InterruptedException at hudson.plugins.buckminster.EclipseBuckminsterBuilder.perform(EclipseBuckminsterBuilder.java:265) at hudson.plugins.templateproject.ProxyBuilder.perform(ProxyBuilder.java:83) at hudson.tasks.BuildStepMonitor$1.perform(BuildStepMonitor.java:19) at hudson.model.AbstractBuild$AbstractRunner.perform(AbstractBuild.java:663) at hudson.model.Build$RunnerImpl.build(Build.java:177) at hudson.model.Build$RunnerImpl.doRun(Build.java:139) at hudson.model.AbstractBuild$AbstractRunner.run(AbstractBuild.java:429) at hudson.model.Run.run(Run.java:1375) at hudson.model.FreeStyleBuild.run(FreeStyleBuild.java:46) at hudson.model.ResourceController.execute(ResourceController.java:88) at hudson.model.Executor.run(Executor.java:145) Caused by: java.lang.InterruptedException at java.lang.Object.wait(Native Method) at java.lang.Object.wait(Object.java:485) at java.lang.UNIXProcess.waitFor(UNIXProcess.java:165) at hudson.Proc$LocalProc.join(Proc.java:319) at hudson.plugins.buckminster.EclipseBuckminsterBuilder.perform(EclipseBuckminsterBuilder.java:261) ... 10 more I have increased the build time to 30 minutes and the tests finish successfully. The reason for the long test run is, that the default timeout is configured to be high. We might wanna change this in the tests. I have downloaded the latest zookeeper and it seems to be working fine, without any visible delays. This test is not extensive since it is hard for me to consume a development build ECF in our official build. However, I have updated to the latest ECF in my Eclipse and the local launch seems to work fine. TBC |