Re: toast table growing indefinitely? Known problems?

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Philip Warner <pjw(at)rhyme(dot)com(dot)au>
Cc: pgsql-hackers(at)postgresql(dot)org
Subject: Re: toast table growing indefinitely? Known problems?
Date: 2002-12-02 05:37:29
Message-ID: 9399.1038807449@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Philip Warner <pjw(at)rhyme(dot)com(dot)au> writes:
> Is there a known problem with TOAST tables growing indefinitely in 7.2?

No. The index on the toast table may well bloat, because it's tracking
a moving range of toast-item OIDs. But the table itself should be okay
so long as it's vacuumed regularly.

If you're seeing bloat while using plain (not full) vacuums, my guess
would be that you need to enlarge the FSM parameters in postgresql.conf.

> On a separate issue, we have seen toast indexes growing indefinitely on 7.3
> despite VACUUM FULL. Attempting a reindex results in:

> # reindex table tt;
> WARNING: table "tt" wasn't reindexed
> REINDEX

AFAIK that will only reindex tt's own indexes. To reindex the toast
table would require naming same, with something like:

regression=# reindex table pg_toast.pg_toast_1675403;
REINDEX

or you could specify the index:

regression=# reindex index pg_toast.pg_toast_1675403_index;
REINDEX

The number that appears in these names is the OID of the table owning
the toast table (ie, tt's oid).

regards, tom lane

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Philip Warner 2002-12-02 06:37:09 Re: toast table growing indefinitely? Known
Previous Message Joe Conway 2002-12-02 03:36:45 Re: 7.4 Wishlist