ONDBSPACEDOWN config parameter

Recently, I discovered the importance of the ONDBSPACEDOWN parameter in ver 7.1. When performing a warm restore on a non-critical down chunk, our users began complaining of problems. We found the hard way that our ONDBSPACEDOWN was set to 2 which is "wait". The system hung with a blocked checkpoint. We had to bring the system down and change the parameter to 0 so that we could re-start the warm restore. This cost a couple of hours of down time. Questions - 1. What is the ramification of having ONDBSPACEDOWN = 0 when using fragmented tables? 2. If a chunk containing a table fragment goes down, could users still query the fragmented table and get unpredictable query results? Please help. Thanks in advance, James May mayj@fdhc.state.fl.us