Date: Tue, 27 May 2014 20:00:01 GMT From: Mark Felder <feld@freebsd.org> To: freebsd-ports-bugs@FreeBSD.org Subject: Re: ports/187286: [bsd.default-versions.mk] [patch] Default PostgreSQL version needs to be updated to 9.3 Message-ID: <201405272000.s4RK01s6003087@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR ports/187286; it has been noted by GNATS.
From: Mark Felder <feld@freebsd.org>
To: Francois Tigeot <ftigeot@wolfpond.org>
Cc: bug-followup@FreeBSD.org, Gavin Atkinson <gavin@FreeBSD.org>,
Baptiste Daroussin <bapt@FreeBSD.org>, mat@freebsd.org
Subject: Re: ports/187286: [bsd.default-versions.mk] [patch] Default
PostgreSQL version needs to be updated to 9.3
Date: Tue, 27 May 2014 14:57:42 -0500
On 2014-05-23 07:51, Francois Tigeot wrote:
> On Sun, May 18, 2014 at 08:18:05PM -0500, Mark Felder wrote:
>>=20
>> On May 18, 2014, at 16:25, Francois Tigeot <ftigeot@wolfpond.org>=20
>> wrote:
>>=20
>> > There may be some performance differences between Postgres 9.3 and
>> > Postgres 9.2 on FreeBSD but it's nothing compared to the gap between
>> > FreeBSD and Linux.
>>=20
>> This is true, but we may not want to widen the performance gap=20
>> experienced by the general userbase.
>>=20
>> "With 9.2, up to 17 % more TPS could be achived (for 32 clients)."=20
>> -girgen@
>=20
> This may be true but has nothing to do with the matter at hand.
>=20
> Postgres 9.0 and 9.1 are years old, they are useless.
> 9.4 is currently in beta and the 9.5 development cycle has already=20
> started.
>=20
> I'm currently at PGCon and one of the frequent complaints I have been=20
> hearing
> about is operating system vendors continuing to ship obsolete versions.
>=20
> By refusing to update the default version of PostgreSQL used in ports=20
> to
> something relatively recent, you're effectively rendering FreeBSD less
> attractive as a platform.
CC'ing mat@
Francois has an excellent point. Maybe we should just do the exp-run on=20
9.3 and be done with it? Let the pg guys and our guys figure out the=20
performance impact (which might be false as tests were not done with=20
*remote* clients) and at least keep our platform attractive to postgres=20
users.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?201405272000.s4RK01s6003087>