On Jan 27, 2008 6:45 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Yeah, Rae Steining was complaining to me about that off-list a few weeks
> ago. The whole syncscan behavior risks breaking many apps that "always
> worked before", even if they were disregarding the letter of the SQL spec.
>
> Maybe a GUC variable to enable/disable syncscan?
I'm not sure it's really a good reason for that because it's just a
matter of time for them to be broken anyway.
But it seems at least a good idea to have a way to build reproducible
test cases on production boxes without being perturbed by the other
scans running. Would it need a restart and be a global GUC variable or
could it be set temporarily per session?
--
Guillaume
