After a diagnostic server is restarted, connections turn white or greyed out. None of the connections are not able to make a connection to the monitored instances.
Orphan processes are still lingering and were not properly ended
WORKAROUND:
We suspect what is happening is that the Diagnostic Server is monitoring too many servers that the files are not closing correctly and has a problem writing and reading from the files.
We see errors like the following from the adk logs:
07 May 2011 05:55:27,278 WARN [ShutdownFromSCM] Agent - Not all instances were shut down; only 138 of 165
07 May 2011 05:55:27,293 ERROR [Native Collector Shutdown] OutOfProcessCollector - stop(): Unexpected exception: java.io.IOException: Error writing to server
07 May 2011 05:55:32,309 ERROR [Queue Manager] ThreadServiceImpl - Failed to execute
08 May 2011 05:55:38,428 INFO [Launcher] Launcher - Diagnostic Server Starting...
08 May 2011 05:55:38,428 INFO [Launcher] Launcher - Framework build: 7.0.0 (4509) on Apr 23 2010 16:53
09 May 2011 05:55:23,679 WARN [ShutdownFromSCM] Agent - Not all instances were shut down; only 144 of 165
09 May 2011 05:55:23,710 ERROR [Native Collector Shutdown] OutOfProcessCollector - stop(): Unexpected exception: java.io.IOException: Error writing to server
09 May 2011 05:55:29,554 INFO [ShutdownFromSCM] Launcher - Shutdown completed
If you would like to test and not lose all of your connections, you can back up C:\Documents and Settings\<your uname>\Application Data\Quest Software\Shared\Profiles, then just recreate one connection and see if it connects.
