"scott.marlowe" <scott.marlowe@ihs.com> writes:
> I'd be glad to test it, but we don't have any issues with fk deadlocks
> since our load is 99% read, 1% write, and most of the tables with fks on
> them only have a handful of writers, so any testing I would do would
> probably just be the "we used it in production and it didn't die" kind of
> testing.
That's what I'm looking for mostly: that it does not have any adverse
side-effects.
regards, tom lane
