Subject: Re: pgsql: Add support for temporary replication slots
To: Tom Lane <tgl@sss.pgh.pa.us>
References: <E1cGRP8-0003EH-So@gemulon.postgresql.org>
 <27274.1481557380@sss.pgh.pa.us> <27833.1481558144@sss.pgh.pa.us>
 <62935e6f-4f1b-c433-e0fa-7f936a38b3e5@2ndquadrant.com>
 <22907.1481571177@sss.pgh.pa.us> <15713.1481585953@sss.pgh.pa.us>
Cc: Peter Eisentraut <peter_e@gmx.net>, pgsql-committers@postgresql.org
From: Petr Jelinek <petr.jelinek@2ndquadrant.com>
Message-ID: <314cae39-2e7d-f5a3-d226-cfb80cf80481@2ndquadrant.com>
Date: Tue, 13 Dec 2016 01:05:08 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101
 Thunderbird/45.5.1
MIME-Version: 1.0
In-Reply-To: <15713.1481585953@sss.pgh.pa.us>
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Precedence: bulk
Sender: pgsql-committers-owner@postgresql.org

On 13/12/16 00:39, Tom Lane wrote:
> I wrote:
>> Petr Jelinek <petr.jelinek@2ndquadrant.com> writes:
>>> Attached is the fixed test using the loop from test_extensions (and yes
>>> I would have missed the pg_stat_clear_snapshot() call without the hint,
>>> thanks :) ).
> 
>> Pushed, thanks.
> 
> Hm, buildfarm says this didn't fix it.  Where exactly does the dropping
> of the slot happen ... is it not complete by the time the backend exits?
> 

Yes, I've seen. The cleanup of slots is done in ProcKill(), I wonder,
since it's on_shmem_exit hook, and the pgstat_beshutdown_hook is as
well, maybe the pgstat_beshutdown_hook is called before ProcKill and the
query is lucky to hit right in between.

I am not quite sure what would be the best way to work around that. We
could wait for the slot to disappear instead of trying to drop it, but
then the test would hang on failure.

-- 
  Petr Jelinek                  http://www.2ndQuadrant.com/
  PostgreSQL Development, 24x7 Support, Training & Services


-- 
Sent via pgsql-committers mailing list (pgsql-committers@postgresql.org)
To make changes to your subscription:
http://www.postgresql.org/mailpref/pgsql-committers