Re: Hot standby and b-tree killed items - Mailing list pgsql-hackers

From Simon Riggs
Subject Re: Hot standby and b-tree killed items
Date
Msg-id 1230718360.4032.39.camel@ebony.2ndQuadrant
Whole thread Raw
In response to Re: Hot standby and b-tree killed items  (Heikki Linnakangas <heikki.linnakangas@enterprisedb.com>)
List pgsql-hackers
On Tue, 2008-12-30 at 18:31 +0200, Heikki Linnakangas wrote:
> Simon Riggs wrote:
> > (a) always ignore LP_DEAD flags we see when reading index during
> > recovery.
> 
> This sounds simplest, and it's nice to not clear the flags for the 
> benefit of transactions running after the recovery is done.

Agreed.

(Also: Transaction hint bits are always set correctly, because we would
only ever see a full page write with hints set after the commit/abort
record was processed. So I continue to honour transaction hint bit
reading and setting during recovery).

> You have to be careful to ignore the flags in read-only transactions 
> that started in hot standby mode, even if recovery has since ended and 
> we're in normal operation now.

Got that.

I'm setting ignore_killed_tuples = false at the start of any index scan
during recovery. And kill_prior_tuples is never set true when in
recovery. Both measures are AM-agnostic.

-- Simon Riggs           www.2ndQuadrant.comPostgreSQL Training, Services and Support



pgsql-hackers by date:

Previous
From: Laurent Coustet
Date:
Subject: Re: parallel restore
Next
From: Magnus Hagander
Date:
Subject: Re: version() output vs. 32/64 bits