Re: pg_stats_recovery view

Alvaro Herrera <alvherre@commandprompt.com>

From: Alvaro Herrera <alvherre@commandprompt.com>
To: Jaime Casanova <jaime@2ndquadrant.com>
Cc: Magnus Hagander <magnus@hagander.net>, Fujii Masao <masao.fujii@gmail.com>, Bernd Helmle <mailings@oopsware.de>, Pg Hackers <pgsql-hackers@postgresql.org>
Date: 2012-02-14T11:40:35Z
Lists: pgsql-hackers
Excerpts from Jaime Casanova's message of mar feb 14 04:10:58 -0300 2012:
> On Thu, Feb 2, 2012 at 2:32 AM, Magnus Hagander <magnus@hagander.net> wrote:
> >
> > I haven't looked through the code in detail, but one direct comment:
> > do we really need/want to send this through the stats collector? It
> > will only ever have one sender - perhaps we should just either store
> > it in shared memory or store it locally and only send it on demand?
> >
> 
> fyi, i intend to send a reworked patch later today, it will store the
> info locally and send it on demand.
> about the _short_desc functions, i added that because i wanted to
> understand what was happening during recovery and the wal_record_type
> (xl_info) being a number is not that clear

Maybe it'd be clearer if you display it in hex and filter out just the
bits that are interesting for this use?  IIRC xl_info carries some other
bits than the ones to identify the record type, which could be
confusing.

-- 
Álvaro Herrera <alvherre@commandprompt.com>
The PostgreSQL Company - Command Prompt, Inc.
PostgreSQL Replication, Consulting, Custom Development, 24x7 support