About the ProcessReconnect Barrier

The ProcessReconnect barrier prevents the broker from starting managed processes until any running local processes have had time to reconnect. This ensures that the processes will be accurately counted and protects against database automation starting extra processes. When the ProcessReconnect barrier is tripped a message like the following will be written to the broker's log.

java.io.IOException: Timed out (3 SECONDS) waiting to be ready to start managed 
   process: java.io.IOException: ProcessBarrier[ProcessReconnect] 
   on [ManagedOnly] not ready, see documentation for details
   at com.nuodb.agent.ProcessBarrierManager.waitCanStartProcess
      (ProcessBarrierManager.java:63)
   at com.nuodb.agent.service.DescriptionService.createDatabase
      (DescriptionService.java:368)
   at com.nuodb.agent.service.DescriptionService.handleRequest
      (DescriptionService.java:238)
    ...

If you see this message in a broker's log, no action needs to be taken. If you encounter it while trying to create a managed database, wait ten (10) seconds and retry database creation.

Upon startup the broker reports any local processes that it finds, including itself, in its log. TEs and SMs are shown as "nuodb" (with ".exe" on WIndows). Following is an agent startup ps log entry:

2015-01-15T13:50:46.244-0500 INFO SigarContainerImpl.printNuodbProcStats (main) 
   nuodb ps
79908    jdoe    15:14    2.4G     21M    -    R    0:0    
   /var/opt/nuodb/bin/nuodb --connect-key 4701283666350936203
79907    jdoe    15:14    2.4G     21M    -    R    0:0    
   /var/opt/nuodb/bin/nuodb --connect-key 7842101632350475923
79906    jdoe    15:14    2.8G     48M    -    R    0:0    
   java:com.nuodb.manager.NuoManager
79905    jdoe    15:14    2.8G     88M    -    R    0:2    
   java:com.nuodb.agent.NuoAgent
...
2015-01-15T13:50:51.363-0500 INFO SigarContainerImpl$ReconnectTimer.run 
   (Thread-9) finished waiting for engines to reconnect