From: | Warren Little <warren(dot)little(at)meridiascapital(dot)com> |
---|---|
To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
Cc: | pgsql-admin(at)postgresql(dot)org |
Subject: | Re: questions on toast tables |
Date: | 2006-04-30 16:55:58 |
Message-ID: | 1146416158.2425.0.camel@wjlnotebook |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
Tom,
thanks much for your help, the cluster command did the trick.
fyi running 8.1.2
On Sat, 2006-04-29 at 14:48 -0400, Tom Lane wrote:
> Warren Little <warren(dot)little(at)meridiascapital(dot)com> writes:
> > Could this be the reference to the toast table that is preventing the
> > vacuum from deleting the toast data? And what purges "dropped" columns
> > if not a full vacuum.
>
> Actually, the way that toast references work is that they'll go away at
> the next update of the row containing the reference. The reason you've
> still got a pile of unremovable toast data is evidently that a lot of
> the parent table's rows have remained untouched since the wide bytea
> column existed. (We choose not to do this housekeeping immediately
> during DROP COLUMN, but to defer it until the next row update.)
>
> One way to clean up the junk would be to do a trivial full-table update
> ("UPDATE foo SET f1 = f1") and then VACUUM FULL, but there are other
> ways that are more efficient. If you're using a PG version released
> within the last year, CLUSTER will do the job nicely.
>
> regards, tom lane
--
Warren Little
Chief Technology Officer
Meridias Capital Inc
1018 W Atherton Dr
SLC, UT 84123
ph 866.369.7763
From | Date | Subject | |
---|---|---|---|
Next Message | Koen Martens | 2006-05-01 12:50:54 | Re: Socket Timeouts and fatal errors (please help) |
Previous Message | Tom Lane | 2006-04-29 19:13:44 | Re: WAL recovery question - 0000001.history |