Thank you for your suggestion. I know that our configuration is not optimal and I will change this. But my understanding was, that this is only a performance problem and not something that can cause Firebird to hang. Or is this a known behavior?
<SL> The problem is in the definition of “hang”. <SL> Does Firebird hang, ie. plain stop? No <SL> Can the deadlock scans and lock contention (see Mutex Wait) make Firebird appear to be hung/running really slow? Yes <SL> FYI, we are considering changing from Classic to SuperClassic to eliminate the overhead (10-15%) that the external lock manager adds. Like you, our databases can be very large (300GB) and have a large number of connection (sometimes 800). This may be something that you also evaluate.
