This has me pulling my hair out. We have a workflow, hosted as a WCF service, which makes a call to another WCF service which then calls a stored procedure. Store procedure calls a merge, then iterates through a cursor that calls another sproc. The cursor count is the same as the source count in the merge. If the source count is high (~120k), the sproc never returns. Disk activity and CPU utilization is nill, and memory is not being taxed. If I then call the sproc from SSMS, it completes in about an hour.
We are using a SQLDataAdapter to make the actual call. Is the SDA somehow receiving updates on every cursor iteration and then failing, causing SQL to stall as it waits? Or is something else going on?
I bring up the SDA because I spent the early part of the week tracking down the cause of the workflow faulting, and it turned out it was a repeated ANSI warning message being returned to the SDA and causing an out of memory exception. Which leads me to wonder if there is something else going on under the cover here, causing the problem.
Turns out it was a very bad case of parameter sniffing. Very, very bad case…