Message-ID: <52E4C0ED4C424CF5BFA534D92A8534F9@Gantry>
From: "Joseph Mays" <mays@win.net>
To: <pgsql-admin@postgresql.org>
Subject: Pgpool-II, pgcluster, or pgbouncer?
Date: Thu, 17 Jul 2014 12:53:49 -0400
MIME-Version: 1.0
Content-Type: multipart/alternative;
 boundary="----=_NextPart_000_043C_01CFA1BE.278DDB20"
Importance: Normal
Precedence: bulk
Sender: pgsql-admin-owner@postgresql.org

This is a multi-part message in MIME format.

------=_NextPart_000_043C_01CFA1BE.278DDB20
Content-Type: text/plain;
	charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

I have a site that has a postgres database server. 24 hour availability =
is pretty critical to them, so I want to turn it into a redundant =
cluster for high availability. I currently have a streaming backup =
server, but we want to be able to make either server capable of =
accepting writes in the case one fails. The problem with streaming =
backup is that we can promote the backup when the master fails, but when =
the master comes backup you basically have to rebuild the entire =
database and reconfigure postgres on it to either make it the master or =
make it into a new streaming backup. We want something that=E2=80=99s =
much more dynamic, so writes can go either direction in the event of one =
or the failing, and so that things will automatically resync when a =
failed server in the cluster comes back online.

Given these goals =E2=80=93 pgpool-II, pgcluster, pgbouncer? Some other =
tool that I haven=E2=80=99t even considered?


------=_NextPart_000_043C_01CFA1BE.278DDB20
Content-Type: text/html;
	charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<HTML><HEAD></HEAD>
<BODY dir=3Dltr>
<DIV dir=3Dltr>
<DIV style=3D"FONT-SIZE: 12pt; FONT-FAMILY: 'Calibri'; COLOR: #000000">
<DIV>I have a site that has a postgres database server. 24 hour =
availability is=20
pretty critical to them, so I want to turn it into a redundant cluster =
for high=20
availability. I currently have a streaming backup server, but we want to =
be able=20
to make either server capable of accepting writes in the case one fails. =
The=20
problem with streaming backup is that we can promote the backup when the =
master=20
fails, but when the master comes backup you basically have to rebuild =
the entire=20
database and reconfigure postgres on it to either make it the master or =
make it=20
into a new streaming backup. We want something that=E2=80=99s much more =
dynamic, so=20
writes can go either direction in the event of one or the failing, and =
so that=20
things will automatically resync when a failed server in the cluster =
comes back=20
online.<BR></DIV>
<DIV>Given these goals =E2=80=93 pgpool-II, pgcluster, pgbouncer? Some =
other tool that I=20
haven=E2=80=99t even considered?</DIV>
<DIV>&nbsp;</DIV>
<DIV>&nbsp;</DIV></DIV></DIV></BODY></HTML>

------=_NextPart_000_043C_01CFA1BE.278DDB20--