Re: adding a new column in IDENTIFY_SYSTEM - Mailing list pgsql-hackers

From Andres Freund
Subject Re: adding a new column in IDENTIFY_SYSTEM
Date
Msg-id 201105040132.01774.andres@anarazel.de
Whole thread Raw
In response to adding a new column in IDENTIFY_SYSTEM  (Jaime Casanova <jaime@2ndquadrant.com>)
Responses Re: adding a new column in IDENTIFY_SYSTEM  (Jaime Casanova <jaime@2ndquadrant.com>)
List pgsql-hackers
Hi,

On Wednesday, May 04, 2011 12:39:49 AM Jaime Casanova wrote:
> I want to propose the addition of a new field in IDENTIFY_SYSTEM:
> xlogversion, which will carry XLOG_PAGE_MAGIC from primary.
> The idea of sending that info is to allow us to know if the xlog page
> version of two different major versions are compatible or not.
> Currently pg_upgrade requires the primary to be taken down, and then
> re-run all base backups for each standby which is a lot of work and
> doesn't sound like "online upgrade".

> I want to add the field now to make the protocol stable, also because
> when we connect to start replication we check for the number of fields
> retrieved from IDENTIFY_SYSTEM so if we add it in 9.2 we will be
> unable to do this until 9.3 (when both releases agree about the number
> of fields returned).
I can't see xlog replication working between major versions. How should it 
handle the differing catversion of the system catalogs? You sure can't 
replicate any of the system catalog changes.

Even if that weren't the case I don't like the idea of xlog improvements 
weighted against the problem of breaking compatibility between major releases.

Or are you proposing not invalidating the base backup? I can't really see that 
working either. You would need to copy over the catalog from the master in the 
exactly right version to the restartpoint on the standby.

Andres


pgsql-hackers by date:

Previous
From: Jaime Casanova
Date:
Subject: adding a new column in IDENTIFY_SYSTEM
Next
From: Josh Berkus
Date:
Subject: Re: Extreme bloating of intarray GiST indexes