Subject: Re: [PATCH] Stop ALTER SYSTEM from making bad assumptions
To: Tom Lane <tgl@sss.pgh.pa.us>, Tomas Vondra <tomas.vondra@2ndquadrant.com>
Cc: Amit Kapila <amit.kapila16@gmail.com>, Stephen Frost
 <sfrost@snowman.net>, Robert Haas <robertmhaas@gmail.com>,
 "pgsql-hackers@postgresql.org" <pgsql-hackers@postgresql.org>
References: 
 <CA+TgmobVjrvVQB_gVytgBcPRmBvptN3wHQy5T6=dYCDZ5n5MGA@mail.gmail.com>
 <20372.1561134638@sss.pgh.pa.us> <20190621164028.GO2480@tamriel.snowman.net>
 <21353.1561136112@sss.pgh.pa.us>
 <CA+TgmoZVx-vco-39EN9MxO9BaG-oVY+NuDeZPGdS=_bX5VH70w@mail.gmail.com>
 <20190624185647.GW2480@tamriel.snowman.net> <15032.1561403529@sss.pgh.pa.us>
 <20190624192014.GB2480@tamriel.snowman.net>
 <56aff4f3-dec8-f462-7feb-0938411b04a6@2ndquadrant.com>
 <CAA4eK1+JMm=tvx720y6jhD8pdzt=HdS1-TwQwkARib2pwK8EvQ@mail.gmail.com>
 <20190802222254.qsvqjtkwte464lu3@development>
 <31697.1564784856@sss.pgh.pa.us>
From: Ian Barwick <ian.barwick@2ndquadrant.com>
Message-ID: <c87c6599-8f60-754f-620d-c30faa0dceea@2ndquadrant.com>
Date: Sat, 3 Aug 2019 07:39:27 +0900
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.7.0
MIME-Version: 1.0
In-Reply-To: <31697.1564784856@sss.pgh.pa.us>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Language: en-US
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 8/3/19 7:27 AM, Tom Lane wrote:
> Tomas Vondra <tomas.vondra@2ndquadrant.com> writes:
>> There seems to be a consensus that this this not a pg_basebackup issue
>> (i.e. duplicate values don't make the file invalid), and it should be
>> handled in ALTER SYSTEM.
> 
> Yeah.  I doubt pg_basebackup is the only actor that can create such
> situations.
> 
>> The proposal seems to be to run through the .auto.conf file, remove any
>> duplicates, and append the new entry at the end. That seems reasonable.
> 
> +1
> 
>> There was a discussion whether to print warnings about the duplicates. I
>> personally see not much point in doing that - if we consider duplicates
>> to be expected, and if ALTER SYSTEM has the license to rework the config
>> file any way it wants, why warn about it?
> 
> Personally I agree that warnings are unnecessary.

Having played around with the pg.auto.conf stuff for a while, my feeling is
that ALTER SYSTEM does indeed have a license to rewrite it (which is what
currently happens anyway, with comments and include directives [1] being silently
removed) so it seems reasonable to remove duplicate entries and ensure
the correct one is processed.

[1] suprisingly any include directives present are honoured, which seems crazy
to me, see: https://www.postgresql.org/message-id/flat/8c8bcbca-3bd9-dc6e-8986-04a5abdef142%402ndquadrant.com

>> The main issue however is that no code was written yet.
> 
> Seems like it ought to be relatively simple ... but I didn't look.

The patch I originally sent does exactly this.

The thread then drifted off into a discussion about providing ways for
applications to properly write to pg.auto.conf while PostgreSQL is not
running; I have a patch for that which I can submit later (though it
is a thing of considerable ugliness).


Regards

Ian Barwick

-- 
  Ian Barwick                   https://www.2ndQuadrant.com/
  PostgreSQL Development, 24x7 Support, Training & Services