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

Re: Disabled features on Hot Standby

From: Robert Haas <robertmhaas(at)gmail(dot)com>
To: Noah Misch <noah(at)leadboat(dot)com>
Cc: Simon Riggs <simon(at)2ndquadrant(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: Disabled features on Hot Standby
Date: 2012-01-14 05:42:02
Message-ID: (view raw, whole thread or download thread mbox)
Lists: pgsql-hackers
On Fri, Jan 13, 2012 at 8:02 PM, Noah Misch <noah(at)leadboat(dot)com> wrote:
> Simon spoke to the FPI side of the question.  For heap pages, the
> TABLESPACE.  For the last, we will have already logged any PD_ALL_VISIBLE bits
> through normal channels.  CLUSTER and VACUUM FULL never set PD_ALL_VISIBLE or
> visibility map bits.  When, someday, they do, we might emit a separate WAL
> record to force the recovery conflict.  However, CLUSTER/VACUUM FULL already
> remove tuples still-visible to standby snapshots without provoking a recovery
> conflict.  (Again only with hot_standby_feedback=off.)

Is the big about CLUSTER/VACUUM FULL a preexisting bug?  If not, why not?

Other than that, it seems like we might be converging on a workable
solution: if hot_standby_feedback=off, disable index-only scans for
snapshots taken during recovery; if hot_standby_feedback=on, generate
recovery conflicts when a snapshot's xmin precedes the youngest xmin
on a page marked all-visible, but allow the use of index-only scans,
and allow sequential scans to trust PD_ALL_VISIBLE.  Off the top of my
head, I don't see a hole in that logic...

Robert Haas
The Enterprise PostgreSQL Company

In response to


pgsql-hackers by date

Next:From: Robert HaasDate: 2012-01-14 05:44:06
Subject: Re: Disabled features on Hot Standby
Previous:From: Robert HaasDate: 2012-01-14 05:27:08
Subject: Re: Disabled features on Hot Standby

Privacy Policy | About PostgreSQL
Copyright © 1996-2017 The PostgreSQL Global Development Group