Weblounge intermittently hangs during startup

The logs always look the same. After "Starting rest publishing service", there is no more output:

2013-08-08 03:43:12 INFO (WebloungeKernel:53) Starting common weblounge services 2013-08-08 03:43:12 INFO (WebloungeKernel:57) Starting job scheduler 2013-08-08 03:43:12 INFO (DispatcherServiceImpl:100) Activating weblounge dispatcher 2013-08-08 03:43:12 INFO (DispatcherServiceImpl:111) Instance name is 'Standalone' 2013-08-08 03:43:12 INFO (SharedHttpContextImpl:113) Enabling requests to protected resources 2013-08-08 03:43:12 INFO (SharedHttpContextImpl:61) Publishing shared http context 2013-08-08 03:43:12 INFO (WebloungeSharedResources:179) Serving shared resources at /weblounge-shared 2013-08-08 03:43:12 INFO (EndpointPublishingService:124) Starting rest publishing service

After attaching the debugger, it turns out that Weblounge is hanging in the following stack:

waiting for: ServiceRegistry (id=65) Object.wait(long) line: not available [native method] ServiceRegistry(Object).wait() line: 502 ServiceRegistry.getService(Bundle, ServiceReference) line: 281 Felix.getService(Bundle, ServiceReference) line: 3568 BundleContextImpl.getService(ServiceReference~~) line: 468 EndpointPublishingService$JSR311AnnotatedServiceListener.serviceChanged(ServiceEvent) line: 332 EndpointPublishingService.activate(ComponentContext) line: 158 ...~~

Fixed by ignoring bundle in "starting" stage during inital scan.

Error still occurs with current develop code-base. Are there any changes in the code associated with this issue elsewhere?

Maybe updating the Felix SCR component would help as well, see https://issues.apache.org/jira/browse/FELIX-2450

Updating Felix SCR to version 1.8.2 seems to help. Instead of blocking the log now occasionally outputs some error messages:

2014-08-20 06:48:27  INFO (WebloungeSharedResources:179) - Serving shared resources at /weblounge-shared
2014-08-20 06:48:27  INFO (EndpointPublishingService:124) - Starting rest publishing service
2014-08-20 06:48:32 ERROR (?:?) - [Weblounge Endpoint Runtime Information(33)] Cannot create component instance due to failure to bind reference endpointpublisher
2014-08-20 06:48:32 ERROR (?:?) - [Weblounge Endpoint Runtime Information(33)] Failed creating the component instance; see log for reason
2014-08-20 06:48:32 ERROR (?:?) - [Weblounge Endpoint Runtime Information(33)] Cannot create component instance due to failure to bind reference endpointpublisher
2014-08-20 06:48:32 ERROR (?:?) - [Weblounge Endpoint Runtime Information(33)] Failed creating the component instance; see log for reason
2014-08-20 06:48:34  INFO (SpringSecurityConfigurationService:221) - Enabling spring security
2014-08-20 06:48:34  INFO (SpringSecurityConfigurationService:234) - Securing the Felix management console

There's an other error message which occurs from time to time:

2014-08-20 10:07:22 ERROR (?:?) - [Weblounge Endpoint Runtime Information(33)] Circular reference detected, getService returning null

But the system is still starting without any problems.

Unfortunately the problem does not seem to be completely gone with the update of Felix SCR to 1.8.2. I can still observe it from time to time.

digitalroastery / weblounge

Weblounge intermittently hangs during startup #290