Re: Replication Across Two Servers? - Mailing list pgsql-general

From Scott Marlowe
Subject Re: Replication Across Two Servers?
Date
Msg-id CAOR=d=3FO=qUH+x0Aefucjy58sjD0m1o_PYQLkGoZOxNStX==A@mail.gmail.com
Whole thread Raw
In response to Re: Replication Across Two Servers?  (Carlos Mennens <carlos.mennens@gmail.com>)
List pgsql-general
On Fri, Nov 4, 2011 at 9:59 AM, Carlos Mennens <carlos.mennens@gmail.com> wrote:
>
> On Fri, Nov 4, 2011 at 11:56 AM, Prashant Bharucha <prashantbharucha@yahoo.ca> wrote:
>>
>> Hi Carlos
>>
>> Use Slony "master to multiple slaves" replication system for PostgreSQL supporting cascading (e.g. - a node can feed
anothernode which feeds another node...) and failover. 
>> http://slony.info/
>>
>
> I'm not sure I see the point of using a third party application to do something PostgreSQL can do natively. Am I
missingsomething here? 

Whether it's "third party" is immaterial really, the real issue is
what are your requirements and which method best meets those
requirements.  For certain more complex replication setups, slony is a
better method.  For instance you can create interesting indexes on a
slony slave that are independent of the master, or create views,
materialized or otherwise on a reporting server and so on.  While
streaming replication is easier to setup and maintain, and generally a
bit  more efficient, it's also got a more limited scope of operation.

Also, if you want to run 8.4 for now, which you've tested against, and
move from 8.4 to 9.1 or 9.2 at a later date, slony is built to do just
that, with running from one major version to another being one of the
things it's really good at.

The real answer then is that it comes down to which meets your
requirements the best.  Both are well tested and supported.

pgsql-general by date:

Previous
From: David Kerr
Date:
Subject: Re: Foreign Keys and Deadlocks
Next
From: Scott Marlowe
Date:
Subject: Re: 9.1 replication on different arch