This is an unusual issue. It seems that there’s a problem with some randomization that we depend on when dealing with multiple threads. The good news is we’re looking at removing the code that’s causing the issue in version 3.4. Hopefully you’re able to wait that long. We have an open issue that’s currently being worked on.
Sorry for the delay in getting back to you, this was an old ticket that was assigned to someone not available when your message came through.
Unfortunately if you are getting this issue, then the only solution is to upgrade past 3.4. We were never able to determine the root cause so the entire query structure and process was rewritten.
There is no other answer, though I am sure you were hoping for something else. You will need a valid license maintenance period from when 3.4 was released.