Horton

Re: Non-reserved replication slots and slot advancing

Michael Paquier <michael@paquier.xyz>

View thread View on postgr.es
From: Michael Paquier <michael@paquier.xyz>
To: Alvaro Herrera <alvherre@2ndquadrant.com>
Cc: Andres Freund <andres@anarazel.de>, Postgres hackers <pgsql-hackers@postgresql.org>, Petr Jelinek <petr@2ndquadrant.com>, Simon Riggs <simon@2ndQuadrant.com>
Date: 2018-07-04T01:57:31Z
Lists: pgsql-hackers 
On Tue, Jul 03, 2018 at 01:51:48PM -0400, Alvaro Herrera wrote:
> On 2018-Jul-03, Andres Freund wrote:
>> Fair enough, but that's what a plain slot allows you as well, pretty
>> fundamentally, no? The precise point at which recycling will be blocked
>> will differer, sure.

ReplicationSlotsComputeRequiredLSN() is careful enough to discard slots
which have their restart_lsn set to InvalidXLogRecPtr, so they are not
accounted within the minimum LSN calculated for segment retention.  Any
fake value added by a user advancing a non-reserved slot is.

At the end, are their any objections into fixing the issue and
tightening the advancing API?
--
Michael

Commits

  1. Block replication slot advance for these not yet reserving WAL