Re: should we add a XLogRecPtr/LSN SQL type?

From: Andres Freund <andres(at)2ndquadrant(dot)com>
To: Michael Paquier <michael(dot)paquier(at)gmail(dot)com>
Cc: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>, Robert Haas <robertmhaas(at)gmail(dot)com>, Simon Riggs <simon(at)2ndquadrant(dot)com>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: should we add a XLogRecPtr/LSN SQL type?
Date: 2014-02-01 15:07:45
Message-ID: 20140201150745.GC32407@alap3.anarazel.de
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 2014-02-02 00:04:41 +0900, Michael Paquier wrote:
> On Fri, Dec 13, 2013 at 2:47 AM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> > Andres Freund <andres(at)2ndquadrant(dot)com> writes:
> >> On 2013-12-12 11:55:51 -0500, Tom Lane wrote:
> >>> I'm not, however, terribly thrilled with the suggestions to add implicit
> >>> casts associated with this type. Implicit casts are generally dangerous.
> >
> >> It's a tradeof. Currently we have the following functions returning LSNs
> >> as text:
> >> * pg_current_xlog_location
> >> * pg_current_xlog_insert_location
> >> * pg_last_xlog_receive_location
> >> * pg_last_xlog_replay_location
> >> one view containing LSNs
> >> * pg_stat_replication
> >> and the following functions accepting LSNs as textual paramters:
> >> * pg_xlog_location_diff
> >> * pg_xlogfile_name
> >
> >> The question is how do we deal with backward compatibility when
> >> introducing a LSN type? There might be some broken code around
> >> monitoring if we simply replace the type without implicit casts.
> >
> > Given the limited usage, how bad would it really be if we simply
> > made all those take/return the LSN type? As long as the type's
> > I/O representation looks like the old text format, I suspect
> > most queries wouldn't notice.

I've asked around inside 2ndq and we could find one single problematic
query, so it's really not too bad.

> Are there some plans to awaken this patch (including changing the
> output of the functions of xlogfuncs.c)? This would be useful for the
> differential backup features I am looking at these days. I imagine
> that it is too late for 9.4 though...

I think we should definitely go ahead with the patch per-se, we just
added another system view with lsns in it... I don't have a too strong
opinion whether to do it in 9.4 or 9.5. It seems fairly low impact to
me, and it's an old patch, but I personally don't have the tuits to
refresh the patch right now.

Greetings,

Andres Freund

--
Andres Freund http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training & Services

In response to

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Andres Freund 2014-02-01 15:44:55 IndexBuildHeapScan doesn't use page at a time mode
Previous Message Michael Paquier 2014-02-01 15:04:41 Re: should we add a XLogRecPtr/LSN SQL type?