Resolution:
This article describes tips and techniques which can be used to troubleshoot BW engine hang issues:

1). Check the BW engine log file. If the engine has thrown some exceptions before going into a hang state (or the job has gone into a hang sate) and it is captured in the log file, this information may help in diagnosing the problem.

2). Get the list of the "Active Processes" from the Admin GUI and see at which activity the jobs struck/hanging. By default, the engine has 8 worker threads and if there are at least 8 jobs stuck in some activity the engine might appear to be a in a hung state as it cannot accept anymore jobs.

3). One of the most common occurrence is the JDBC activities hanging without timing out. Refer to KB 42645 for more details.

4). It is also important to note that if "FlowLimit" has been set to some value, say 'n'. Once the Flowlimit is reached the starter is disabled and the engine will only accept new jobs once the engine processed half of the jobs. It is important to check the process starter configuration and the engine tuning parameters so that the configuration mistake does not cause this issue.

5). Find out if just one starter is not accepting jobs or all the starters are not accepting jobs.

6). Assuming the issue is reproducible, run the engine from the command line and not from the Administrator GUI. Make sure to redirect the console (stdout and stderr) output to a file. If there are any errors/exceptions thrown on the console they will get captured. Checking for errors /exceptions in tsm.log can also help. Additionally a Java thread dump (refer KB 28813, 40847) can point to threads which are hung.

7). If the starter is JMS or RV it may be helpful if some JMS or RV tracing is enabled to identify the problem. For example, with JMS we can check the connections, queues/topics and durables.