Subject: Re: ALTER TABLE ... SET STORAGE does not propagate to indexes
From: Peter Eisentraut <peter.eisentraut@2ndquadrant.com>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
References: <20200421235636.GA29434@alvherre.pgsql>
 <8be09419-ed3d-93f9-9c29-2824b44a2c56@2ndquadrant.com>
Cc: Kuntal Ghosh <kuntalghosh.2007@gmail.com>,
 pgsql-hackers <pgsql-hackers@postgresql.org>
Organization: 2ndQuadrant
Message-ID: <1e19aeb2-c300-a8a0-e828-9090671eb428@2ndquadrant.com>
Date: Wed, 6 May 2020 16:37:55 +0200
User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.14; rv:52.0)
 Gecko/20100101 Thunderbird/52.9.1
MIME-Version: 1.0
In-Reply-To: <8be09419-ed3d-93f9-9c29-2824b44a2c56@2ndquadrant.com>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 2020-04-22 16:26, Peter Eisentraut wrote:
> On 2020-04-22 01:56, Alvaro Herrera wrote:
>> I'm surprised that this hasn't applied yet, because:
>>
>> On 2020-Apr-09, Peter Eisentraut wrote:
>>
>>> One thing to remember is that the current situation is broken.  While you
>>> can set index columns to have different storage than the corresponding table
>>> columns, pg_dump does not preserve that, because it dumps indexes after
>>> ALTER TABLE commands.  So at the moment, having these two things different
>>> isn't really supported.
>>
>> So I have to ask -- are you planning to get this patch pushed and
>> backpatched?
> 
> I think I should, but I figured I want to give some extra time for
> people to consider the horror that I created in the test_decoding tests.

OK then, if there are no last-minute objects, I'll commit this for the 
upcoming minor releases.

This is the patch summary again:

Date: Thu, 9 Apr 2020 14:10:01 +0200
Subject: [PATCH v3] Propagate ALTER TABLE ... SET STORAGE to indexes

When creating a new index, the attstorage setting of the table column
is copied to regular (non-expression) index columns.  But a later
ALTER TABLE ... SET STORAGE is not propagated to indexes, thus
creating an inconsistent and undumpable state.

-- 
Peter Eisentraut              http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services