From: | "Scott Marlowe" <scott(dot)marlowe(at)gmail(dot)com> |
---|---|
To: | "BRAHMA PRAKASH TIWARI" <prakashr2n5(at)gmail(dot)com> |
Cc: | "Francesco Andreozzi" <francesco(dot)andreozzi(at)gamestorm(dot)it>, pgsql-admin(at)postgresql(dot)org |
Subject: | Re: tuning tips, speed problem |
Date: | 2009-01-19 15:54:30 |
Message-ID: | dcc563d10901190754o4c8f74eeidad1b52cf2e73356@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
On Mon, Jan 19, 2009 at 5:35 AM, BRAHMA PRAKASH TIWARI
<prakashr2n5(at)gmail(dot)com> wrote:
> Hi Francesco
>
> most probabely this is due to the auto vacuum option on and if auto vacuum
> is on then the in condition of high transaction on database it slows the
> speed of the hole database.Set it off in postgres.conf and vacuum and
> reindex transactional tables manualy with in every two million
> transactions.
> like
> vacuum full <table name>;
> reindex table <table name>;
Except for certain circumstances vacuum full should be avoided and
vacuum (regular) used in its place until it is proven ineffective.
Also, it is usually far better to turn up the
autovacuum_vacuum_cost_delay to 10 or 20 and let autovacuum do its
job. Unless you have a very starved I/O subsystem autovacuum with
cost delay of 20 should have almost no noticeable effect on a
transactional database.
From | Date | Subject | |
---|---|---|---|
Next Message | Jaume Sabater | 2009-01-19 15:56:23 | Re: PostgreSQL monitoring |
Previous Message | Marcelo Martins | 2009-01-19 15:38:36 | PostgreSQL monitoring |