Re: Add system identifier to backup manifest

From: Michael Paquier <michael(at)paquier(dot)xyz>
To: Robert Haas <robertmhaas(at)gmail(dot)com>
Cc: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>, Amul Sul <sulamul(at)gmail(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Add system identifier to backup manifest
Date: 2024-01-18 02:20:17
Message-ID: ZaiK4WD5gPD4MMUX@paquier.xyz
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On Wed, Jan 17, 2024 at 08:46:09AM -0500, Robert Haas wrote:
> On Wed, Jan 17, 2024 at 6:45 AM Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
>> Hmm, okay, but what if I take a full backup from a primary server and
>> later I want an incremental from a standby, or the other way around?
>> Will this prevent me from using such a combination?
>
> The system identifier had BETTER match in such cases. If it doesn't,
> somebody's run pg_resetwal on your standby since it was created... and
> in that case, no incremental backup for you!

There is an even stronger check than that at replay as we also store
the system identifier in XLogLongPageHeaderData and cross-check it
with the contents of the control file. Having a field in the backup
manifest makes for a much faster detection, even if that's not the
same as replaying things, it can avoid a lot of problems when
combining backup pieces. I'm +1 for Amul's patch concept.
--
Michael

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Richard Guo 2024-01-18 02:22:24 Re: Strange Bitmapset manipulation in DiscreteKnapsack()
Previous Message torikoshia 2024-01-18 02:15:47 Re: POC PATCH: copy from ... exceptions to: (was Re: VLDB Features)