Skip site navigation (1) Skip section navigation (2)

Re: heap_tuple_needs_freeze false positive

From: Robert Haas <robertmhaas(at)gmail(dot)com>
To: Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>
Cc: Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: heap_tuple_needs_freeze false positive
Date: 2012-02-02 14:30:48
Message-ID: CA+TgmoZJSv3F=6KpmLCe6rPrbu5WygL=S_tkOmiDV_pG_t=MpA@mail.gmail.com (view raw or flat)
Thread:
Lists: pgsql-hackers
On Wed, Feb 1, 2012 at 8:01 PM, Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org> wrote:
> Suggested patch attached.  I'd backpatch this as far as it applies
> cleanly.

This is new code in 9.2, but it's modelled on heap_freeze_tuple(), which is old.

I'm not convinced that it's a bug.  Suppose that xmax is set but is
hinted as invalid.  We process the table and advanced relfrozenxid;
then, we crash.  After recovery, it's possible that the hint bit is
gone (after all, setting hint bits isn't WAL-logged).  Now we're in
big trouble, because the next CLOG lookup on that xmax value might not
happen until it's been reused, and we might get a different answer
than before.

-- 
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company

In response to

Responses

pgsql-hackers by date

Next:From: Marko KreenDate: 2012-02-02 14:30:57
Subject: Re: Speed dblink using alternate libpq tuple storage
Previous:From: Robert HaasDate: 2012-02-02 14:24:08
Subject: Re: freezing multixacts

Privacy Policy | About PostgreSQL
Copyright © 1996-2014 The PostgreSQL Global Development Group