On Dec 19, 2007, at 6:39 PM, Tom Lane wrote:
>> The thing that concerns me is dead tuples on the table_stats
>> table. I
>> believe that every insert of new data in one of the monitored tables
>> will result in an UPDATE of the table_stats table. When thousands
>> ( or millions ) of rows are inserted, the select performance ( even
>> trying with an index ) on table_stats slows down in a hurry.
>
> Yup. FWIW, 8.3's "HOT" tuple updates might help this quite a lot.
> Not a lot to be done about it in 8.0.x though :-(
A work-around others have used is to have the trigger just insert
into a 'staging' table and then periodically take the records from
that table and summarize them somewhere else. You still have a vacuum
concern on the staging table, but the advantage is that you trigger
path is a simple insert instead of an update, which is effectively a
delete and an insert.
This is a case where a cron'd vacuum that runs once a minute is
probably a wise idea.
--
Decibel!, aka Jim C. Nasby, Database Architect decibel@decibel.org
Give your computer some brain candy! www.distributed.net Team #1828