>Michel, a few months ago, you reported an error in a scenario that sounded a little like the one you just described. And I think it was because of a delete occurring while another process was trying to read the same row. Is this new situation anything like the one you had a while back?

No, this one is new. Basically, the process cannot run anymore. And, this is with low traffic on site. So, this robot process cannot execute this delete command anymore and I really do not know what is the trigger. I am pretty sure it is related to this SQL Server log process of logging every transaction in case we need to recovery. So, a delete plus a write in another place somewhere seems to be quite heavy now for this SQL command to execute. I am wondering if a delete in a loop one record at a time would be sufficent to bypass this behavior.