| From: | Neil Conway <neilc(at)samurai(dot)com> | 
|---|---|
| To: | Ruslan A Dautkhanov <rusland(at)scn(dot)ru> | 
| Cc: | pgsql-bugs(at)postgresql(dot)org | 
| Subject: | Re: database caches | 
| Date: | 2002-10-29 05:27:18 | 
| Message-ID: | 87d6pt3jd5.fsf@mailbox.samurai.com | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-bugs | 
Ruslan A Dautkhanov <rusland(at)scn(dot)ru> writes:
>     I have problem in PG performance - all database caches is allocated
>     to one table, which is much greater than all other tables together, but
>     requests to this huge table much rarely, than requests with
>     other tables.
What makes you think that all the cache is allocated for this one
specific table?
Also, caching is done on a page-by-page basis, not per-relation.
>     How can I adjust which tables should be cached more (cache rate), and
>     which less, if this possible?? Or which tables not to cache
>     completely.
The buffer manager decides what to cache by itself; there are a few
knobs you can tweak (shared_buffers, for example), but for the most
part the buffer manager is expected to do its job.
In your specific example, it doesn't sound like there's a need to
manually tweak the buffer manager: the large but infrequently accessed
table will likely not have very many buffers kept in memory.
Cheers,
Neil
-- 
Neil Conway <neilc(at)samurai(dot)com> || PGP Key ID: DB3C29FC
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Vincent Lim | 2002-10-29 07:44:12 | Re: Postgresql 7.2.3 | 
| Previous Message | Ruslan A Dautkhanov | 2002-10-29 04:25:51 | database caches |