Re: Cluster using tablespaces?

From: Rainer Bauer <usenet(at)munnin(dot)com>
To: pgsql-general(at)postgresql(dot)org
Subject: Re: Cluster using tablespaces?
Date: 2007-11-29 00:05:25
Message-ID: epurk3pp6u5sva4vpc07klsfjuug2d8f4k@4ax.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

Tom Lane wrote:

>Rainer Bauer <usenet(at)munnin(dot)com> writes:
>
>> "During the cluster operation, a temporary copy of the table is created that
>> contains the table data in the index order. Temporary copies of each index on
>> the table are created as well."
>
>That's probably a bit misleading. There is no "temporary" copy of the
>table, just the new permanent copy. The document is trying to point out
>to you that the transient disk space requirement will be 2X the table
>size, but maybe we could phrase it better.

Ok, I expected that. Does this work:
ALTER TABLE foo CLUSTER ON index_name SET TABLESPACE new_tablespace

I.e. is the table moved to the other tablespace and clustered at the same time
or are these independant operations?

What I am trying to achieve is cutting down the time the cluster command
takes. I thought the most promising way would be if the new data is written to
different drive.

>For btree indexes, there is a temporary copy of the index data, which
>will go wherever you have arranged for temp files to go. (I think that
>easy user control of this may be new for 8.3, though.)

Could you give me a hint where that would be on Windows? I guess this might be
worth a try since there are a couple of btree indexes in the database.

Rainer

In response to

Responses

Browse pgsql-general by date

  From Date Subject
Next Message Ron Johnson 2007-11-29 00:27:35 Re: PostgresSQL vs. Informix
Previous Message Tom Lane 2007-11-28 23:30:10 Re: Cluster using tablespaces?