Re: reserving space in a rec for future update

From: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>
To: Erik Jones <erik(at)myemma(dot)com>
Cc: Mike Charnoky <noky(at)nextbus(dot)com>, Andrew Sullivan <ajs(at)crankycanuck(dot)ca>, pgsql-general(at)postgresql(dot)org
Subject: Re: reserving space in a rec for future update
Date: 2007-11-14 17:11:08
Message-ID: 20071114171108.GS19014@alvh.no-ip.org
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Erik Jones wrote:
>
> On Nov 14, 2007, at 10:44 AM, Mike Charnoky wrote:
>
>> In this usage scenario, doesn't the new HOT (heap only tuples) feature
>> of PG8.3 help, in terms of the DB requiring less VACUUM maintenance?
>>
>> I am similarly performing a huge number of inserts, followed by a huge
>> number of updates to fill in a few null fields. The data is indexed by
>> insert time. My problem is, selects using that index degrade over time
>> as updates are performed, presumably because data is no longer ordered
>> sequentially across pages after updates are performed. I was hoping
>> that HOT would help here and am actually installing PG8.3 now in order
>> to perform some testing...
>
> Some, what HOT does is keeps index rows from being updated when updates are
> made to column values that aren't indexed. The same insert/delete still
> happens in the table data.

But another thing HOT does is make it possible to vacuum the *page* that
the update is taking place on. So if there are dead tuples that nobody
needs, they can be removed and the new tuple can be placed there.

--
Alvaro Herrera http://www.advogato.org/person/alvherre
"Use it up, wear it out, make it do, or do without"

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Tony Caduto 2007-11-14 17:22:01 PLpgsql debugger question
Previous Message Erik Jones 2007-11-14 17:02:48 Re: reserving space in a rec for future update