Re: BUG #3479: contraint exclusion and locks - Mailing list pgsql-bugs

From Simon Riggs
Subject Re: BUG #3479: contraint exclusion and locks
Date
Msg-id 1185229400.4284.439.camel@ebony.site
Whole thread Raw
In response to Re: BUG #3479: contraint exclusion and locks  (Tiago Daniel Jacobs <tiago@mdtestudio.com.br>)
List pgsql-bugs
On Mon, 2007-07-23 at 11:46 -0300, Tiago Daniel Jacobs wrote:
> Tom and Gregory.  U're right! The problem is that we're using
> constraints for partitioned tables and by definition, a partition
> never, absolutely never, can affect the entire system.
>
Yeh, the problem is that partitioning uses additional information to
avoid reading data. The additional information and the actual data need
to be kept in step, so at some point we may need to re-write that data
and we'll always need a lock to do that.

So whether we use constraints, segment headers or what have you,
there'll still be a need to lock and be locked out.

--
  Simon Riggs
  EnterpriseDB  http://www.enterprisedb.com

pgsql-bugs by date:

Previous
From: Kris Jurka
Date:
Subject: Re: BUG #3480: JDBC driver: getIndexInfo() returns quotes around quoted column names
Next
From: "Jim C. Nasby"
Date:
Subject: Re: BUG #3476: description of root.crt/crl in documentation