From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
---|---|
To: | "scott(dot)marlowe" <scott(dot)marlowe(at)ihs(dot)com> |
Cc: | Chris Miles <chris_pg002(at)psychofx(dot)com>, pgsql-admin(at)postgresql(dot)org |
Subject: | Re: Does VACUUM ever free up any disk space? |
Date: | 2003-09-11 21:02:27 |
Message-ID: | 1380.1063314147@sss.pgh.pa.us |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
"scott.marlowe" <scott(dot)marlowe(at)ihs(dot)com> writes:
> On Thu, 11 Sep 2003, Chris Miles wrote:
>> I've read a lot where people recommend using VACUUM FULL
>> to free up disk space, especially after many updates/inserts.
> A regular vacuum since 7.2 never actually frees up space, it simply marks
> the space in the file as available for reuse. Note that in many
> circumstances this is actually better than freeing up the space, as it
> allows the database to store date without having to extend and possibly
> fragment the table.
Regular vacuum *will* shorten the table's file if (a) there are some
completely-empty pages at the end, and (b) it can get an exclusive lock
on the table without blocking. This might be a relatively rare
condition in a heavily-used table. But "never actually frees up space"
is incorrect.
You're correct that regular vacuum is designed around the idea of
maintaining a steady-state file size rather than trying very hard to
give space back to the OS.
regards, tom lane
From | Date | Subject | |
---|---|---|---|
Next Message | Andrew Biagioni | 2003-09-11 21:11:21 | Cross-DB linking? |
Previous Message | Peralta Miguel-MPERALT1 | 2003-09-11 19:35:56 | unsubscribe |