Gavin Hamill <gdh(at)laterooms(dot)com> writes:
> The table has just under six million rows - should it really be taking
> nearly six minutes to add an index?
Try running it with trace_sort enabled to get more info about where the
time is going.
We've been doing some considerable work on the sorting code in the last
couple months, so 8.2 should be better, but I'd like to verify that
you're not seeing something we don't know about.
> maintenance_work_mem = 262144
Fooling with this might affect the results some.
regards, tom lane
In response to
pgsql-performance by date
|Next:||From: Tom Lane||Date: 2006-03-30 15:24:23|
|Subject: Re: Index scan startup time |
|Previous:||From: Peter Eisentraut||Date: 2006-03-30 15:16:13|
|Subject: Re: Index scan startup time|