Based on this, the resolution proposed at the beginning of the week, has been to use the recycler process option to force it to recycle at every 1000 hits. When doing so, we had zero memory problem and zero VFPOleDb error message. They tried from 100, to 1000, to 1500 and back to 1000. At that interval, this has done the work so far. So, by doing it, we have resolved, at least for so far, our biggest two issues. Well, those were the only ones we had. So, at least, this has helped. I cannot say it will not happen again. We will have to wait a few more days. But, what I have seen so far, is way much better than what we have seen before.

Basically, this could be the way to go to make sure the VFPOleDb works ok with VFP data. That has been a problem since last year. As for the memory problem, the only parallel I could see was that since a few weeks, the increase of hits has happened. So, because more people were using the site, more hits were being done and that could have caused that. The size of the tables have also increased. So, there could be a corelation in here.