Re: pgsql: Must not reach consistency before XLOG_BACKUP_RECORD

From: Andres Freund <andres(at)2ndquadrant(dot)com>
To: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
Cc: Simon Riggs <simon(at)2ndQuadrant(dot)com>, pgsql-committers(at)postgresql(dot)org
Subject: Re: pgsql: Must not reach consistency before XLOG_BACKUP_RECORD
Date: 2012-12-05 16:11:00
Message-ID: 20121205161100.GA20456@awork2.anarazel.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-committers

On 2012-12-05 11:04:39 -0500, Tom Lane wrote:
> Simon Riggs <simon(at)2ndQuadrant(dot)com> writes:
> > Must not reach consistency before XLOG_BACKUP_RECORD
>
> > When waiting for an XLOG_BACKUP_RECORD the minRecoveryPoint
> > will be incorrect, so we must not declare recovery as consistent
> > before we have seen the record. Major bug allowing recovery to end
> > too early in some cases, allowing people to see inconsistent db.
>
> Is this actually a "major bug fix", or a useless redundant test?
> I had thought that that if-statement's check of
> XLogRecPtrIsInvalid(ControlFile->backupStartPoint)
> was sufficient, because that will keep us from declaring consistency
> before we get out of the backup in any case.

I think youre right that its redundant. Adding the check seems to be
sensible from a robustness perspective in the long run, but its
certainly not relevant for the release process.

Greetings,

Andres

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

In response to

Browse pgsql-committers by date

  From Date Subject
Next Message pgsql 2012-12-05 17:44:03 pgsql: Tag refs/tags/REL8_3_22 was created
Previous Message Tom Lane 2012-12-05 16:04:39 Re: pgsql: Must not reach consistency before XLOG_BACKUP_RECORD