On Aug 9, 10:47 pm, [EMAIL PROTECTED] (Tom Lane) wrote:
> Do you have comparable work_mem settings on both machines? Another
> thing to look at, if any of the sort key columns are textual, is whether
> the lc_collate settings are the same.
work_mem is commented out in both postgresql.conf files:
Adam Endicott <[EMAIL PROTECTED]> writes:
> Here's the output from explain analyze.
Wow, so the differential is all in the sort step.
8.2 does have improved sorting code, but I don't think that explains
the difference, especially not for a mere 16000 rows to be sorted.
Do you have comparable wor
Here's the output from explain analyze.
My desktop:
-
Unique (cost=6732.86..7380.50 rows=504 width=677) (actual
time=844.345..1148.705 rows=65 loops=1)
-> Sort (cost=6732.86..6773.34 rows=16191 width=677) (actual
time=844.341..1099.446 rows=16191 loops=1)
Sort Key:
Adam Endicott <[EMAIL PROTECTED]> writes:
> When I run EXPLAIN ANALYZE on this query, it takes something like
> 1200ms on my desktop (Dual 2GHz G5 Mac - 1.5 GB RAM for reference) and
> about 14000ms on the production server (quad processor, 8 GB RAM,
> running Ubuntu). There are about 500 rows in t
I'm having an issue with a specific query, and I don't really know
where to start figuring out what's going on. I'm pretty new to
PostgreSQL in specific, and I'm not much of a database/SQL guru in
general. I've got one query that is consistently taking 10X longer to
run on a production machine than
