Here's a patch for WAL logging tuple freezes in vacuum, per discussion
on pgsql-bugs.

This patch is against CVS head. Should this be backported to stable
branches? I think it should.

After writing the patch, I realized that it needs some thought if 
backported, because WAL records of removing tuples and freezing tuples 
share the same heapam opcode XLOG_HEAP_CLEAN, and are only 
differentiated by setting a flag. If we applied the patch as it is, and 
for some reason someone replayed a WAL log generated by a newer version, 
with the patch, with an older version, without the patch, the older 
version would interpret the freeze WAL records as dead tuple removals, 
and remove live records. I would've liked to give freezing a new opcode, 
  but we've ran out of them (see htup.h).

-- 
    Heikki Linnakangas
    EnterpriseDB   http://www.enterprisedb.com

Responses

• Re: WAL logging freezing at 2006-10-24 12:11:33 from Alvaro Herrera
• Re: WAL logging freezing at 2006-10-24 14:11:43 from Tom Lane

pgsql-hackers by date

Next:	From: Markus Schiltknecht	Date: 2006-10-24 08:26:39
	Subject: Re: Replication documentation addition
Previous:	From: Bruce Momjian	Date: 2006-10-24 04:20:40
	Subject: Replication documentation addition

pgsql-patches by date

Next:	From: Hitoshi Harada	Date: 2006-10-24 09:52:01
	Subject: Re: [PATCHES] smartvacuum() instead of autovacuum
Previous:	From: Magnus Hagander	Date: 2006-10-24 06:05:28
	Subject: Re: BUG #2712: could not fsync segment: Permission