On Thursday November 11 2004 7:28, Ed L. wrote:
> On Thursday November 11 2004 6:26, Ed L. wrote:
> > So I hacked cleanup_thread.c to skip the if-statement that does the
> > vacuum with query3.  Now I'm seeing it do its 2nd round of
> > truncateTable()-copy during the copy set operation under the same
> > backend pid.  I can't recreate this behavior using the same code on a
> > small pgbench database.

Looks like the problem was my application of the copy set patch, not sure 
why.  Removing the copy set patch fixed the problem.

Ed