> Thanks to vlad, he discovered that using the monitoring table is one raison of one bug (lock error), but unfortunatly it's seam not answer all the other bug (index corrupted for exemple) ...

I looking into index corruptions for a few past days and currently my machine is busy with your test :) Bug with monitoring will be fixed soon, we have few possible solutions for it and we are looking for the best one.

Index corruptions is more hard to fix but i hope i'll found a reason soon. Note, i still not sure if indices was corrupted because of unexpected stop of engine process after "invalid lock id" error.

Regards,
Vlad