Hi Igor,

I'm not sure what could cause that. We are basically concatenating the item title and link and using a hashcode of the resulting string as a way to identify if an item already exists or not. It looks like the title and link are identical so it should produce the same hash code and therefore not create a duplicate in the mp_RssFeedEntries table. I would look at the rows there and see if they have different or the same hash code.

Hope it helps,

Joe

Hi Igor,

This "might" be fixed in svn now. I did some thinking about the problem and realized I should have a lock around the code that updates the data cached in the db so that multiple requests at the same time when cache is expired do not step on each other by trying to refresh the data in the db at the same time. That scenario could have resulted in the duplicates in the db if the timing was just right.

Best,

Joe