Hi Viv

I was wondering if you could run your stress test again about the VFPOleDb issue but by adding a SQL in it that would do a write on a table. I have found something weird in regards to that. While I am not able to simulate the problem with my stress test with the new approach which consists of using a global connection which is already opened for every SQL requests, I believe there is an issue when a write is being done when simultaneous hits are in process. I would be interested to hear about it after your test because you have done some interesting tests so far in regards to this issue.