On Tuesday February 24 2004 1:14, Kevin Brown wrote:
>
> One problem I've been running into is the merge join spilling to disk
> because sort_mem isn't big enough. The problem isn't that this is
> happening, it's that I think the planner is underestimating the impact
> that doing this will have on the time the merge join takes. Does the
> planner even account for the possibility that a sort or join will spill
> to disk? Spilling to disk like that will suddenly cause sequential
> reads to perform much more like random reads, unless the sequential
> scans are performed in their entirety between sorts/merges.
How do you know the merge join is spilling to disk? How are you identifying
that? Just assuming from vmstat? iostat?