I was just about to reply to the thread, thanks for asking. Clustering was
key. After rebooting the machine (just to make sure absolutely nothing was
cached), I immediately ran a report on Toronto: 5.25 seconds!
Here's what I did:
1. Created a new set of tables that matched the old set, with statistics
of 1000 on the station and taken (date) columns.
2. Inserted the data from the old hierarchy into the new set, ordered by
station id then by date (same seven child tables as before: one per
- I wanted to ensure a strong correlation between primary key and
3. Added three indexes per table: (a) station id; (b) date taken; and
4. Set the station-taken-category index as CLUSTER.
5. Vacuumed the new tables.
6. Dropped the old tables.
7. Set the following configuration values:
- shared_buffers = 1GB
- temp_buffers = 32MB
- work_mem = 32MB
- maintenance_work_mem = 64MB
- seq_page_cost = 1.0
- random_page_cost = 2.0
- cpu_index_tuple_cost = 0.001
- effective_cache_size = 512MB
I ran a few more reports (no reboots, but reading vastly different data
- Vancouver: 4.2s
- Yellowknife: 1.7s
- Montreal: 6.5s
- Trois-Riviers: 2.8s
No full table scans. I imagine some indexes are not strictly necessary and
will test to see which can be removed (my guess: the station and taken
indexes). The problem was that the station ids were scattered and so
PostgreSQL presumed a full table scan would be faster.
Physically ordering the data by station ids triggers index use every time.
Next week's hardware upgrade should halve those times -- unless anyone has
further suggestions to squeeze more performance out of PG. ;-)
In response to
pgsql-performance by date
|Next:||From: Cédric Villemain||Date: 2010-05-27 07:50:48|
|Subject: Re: shared_buffers advice|
|Previous:||From: 黄永卫||Date: 2010-05-27 07:27:41|
|Subject: About Tom Lane's Xeon CS test case|