Re: could recovery_target_timeline=latest be the default in standby mode?

From: David Steele <david(at)pgmasters(dot)net>
To: Peter Eisentraut <peter(dot)eisentraut(at)2ndquadrant(dot)com>, Michael Paquier <michael(at)paquier(dot)xyz>
Cc: Sergei Kornilov <sk(at)zsrv(dot)org>, pgsql-hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: could recovery_target_timeline=latest be the default in standby mode?
Date: 2018-12-31 11:21:00
Message-ID: 1c55a6a2-f7fa-4a4a-f90b-60fc5315d781@pgmasters.net
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Hi Peter,

On 12/28/18 1:08 PM, Peter Eisentraut wrote:
> On 28/12/2018 00:15, Michael Paquier wrote:
>> Yes, I was also thinking something among those lines, and the patch is
>> a bit confusing by linking standby mode with latest timeline. It
>> seems to me that switching the default value to "latest" at GUC level
>> would be the way to go, instead of picking up the TLI from the control
>> file. Introducing a new value which maps to the current empty value
>> may be useful as well, like "control_file"?
>
> OK, here are patches for this approach:
>
> 1. Add value 'current' for recovery_target_timeline
> 2. Change default of recovery_target_timeline to 'latest'
>
> The first is really a fixup of the recovery.conf-removal patch. In
> <=PG11, you could not explicitly select the current timeline; it was
> only available if you don't mention recovery_target_timeline at all.
> The original patch contained a setting 'controlfile', similar to your
> suggestion, but that sounds a bit low-level implementation detail to me.
> I like the suggestion 'current'.

This patch looks good to me.

> > The second then just changes the GUC default, without any special
> treatment for standby mode.

Yes, that's exactly what I was thinking.

There don't seem to be any tests for recovery_target_timeline=current.
This is an preexisting condition but it probably wouldn't hurt to add one.

Regards,
--
-David
david(at)pgmasters(dot)net

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Eisentraut 2018-12-31 11:23:46 Re: error message when subscription target is a partitioned table
Previous Message David Steele 2018-12-31 11:07:34 Re: Add timeline to partial WAL segments