Re: Cascading Recovery

On Sun, Feb 13, 2011 at 10:35 AM, Selva manickaraja <mavles78(at)gmail(dot)com>wrote:

> Any updates?

You can't use any wal based replication like that. When a node is in
recovery mode, it doesn't log.

--Scottie

>
>
> ---------- Forwarded message ----------
> From: Selva manickaraja <mavles78(at)gmail(dot)com>
> Date: Sat, Feb 12, 2011 at 1:43 AM
> Subject: Re: [ADMIN] Cascading Recovery
> To: Guillaume Lelarge <guillaume(at)lelarge(dot)info>
> Cc: pgsql-admin(at)postgresql(dot)org
>
>
> If not with Streaming Replication, what about Continous Archiving Mode?
>
>
> On Fri, Feb 11, 2011 at 3:35 PM, Guillaume Lelarge <guillaume(at)lelarge(dot)info
> > wrote:
>
>> Le 11/02/2011 05:43, Selva manickaraja a écrit :
>> > We are considering on using a primary and 2 secondary machines. But our
>> plan
>> > is to cascade the recovery as follows:
>> >
>> > 1. Primary Machine-A (read and write operation)
>> > 2. Secondary Machine-B (recover from Machine A)
>> > 3. Secondary Machine-C (recover from Machine B)
>> >
>> > We are planning this kind of recovery as we want to phase out recovery
>> > interval between C and B 30 minutes.
>> >
>> > Can this be achieved by PostgreSQL.
>> >
>>
>> Not with Streaming Replication. But you can achieve this with a
>> different replication engine, like Slony.
>>
>>
>> --
>> Guillaume
>> http://www.postgresql.fr
>> http://dalibo.com
>>
>
>
>