Hello KaiGain-san,
On 08/19/2015 03:19 PM, Kohei KaiGai wrote:
> Unless we have no fail-safe mechanism when planner estimated too
> large number of tuples than actual needs, a strange estimation will
> consume massive amount of RAMs. It's a bad side effect.
> My previous patch didn't pay attention to the scenario, so needs to
> revise the patch.
I agree we need to put a few more safeguards there (e.g. make sure we
don't overflow INT when counting the buckets, which may happen with the
amounts of work_mem we'll see in the wild soon).
But I think we should not do any extensive changes to how we size the
hashtable - that's not something we should do in a bugfix I think.
regards
--
Tomas Vondra http://www.2ndQuadrant.com
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services