[pgpool-general: 3659] Re: postgresql and pgpool - how best to set up a no data loss solution

[Derek]

Could setting up a pg_xlog folder synchronisation between the master and standby help with reducing the risk of data loss? Or would this cause potential data corruption?

Kind regards,
Derek.

> On 27 Apr 2015, at 11:02 am, Christophe Pettus <xof at thebuild.com> wrote:
> 
> 
>> On Apr 26, 2015, at 5:59 PM, Derek <derek_kouch at yahoo.com.au> wrote:
>> 
>> My question is how can I set up pgpool and / or postgresql so that there is no data loss when a disaster occurs. ie. all committed transactions are sent and loaded at the standby database server as part of the failover.
>> 
>> I want to use an asynchronous connection between the master and standby database servers.
> 
> This isn't directly related to pgpool as such, but there's a conflict between your two requirements above: You can have zero committed transaction loss, or you can have asynchronous replication, but you can't have both.  With asynchronous replication, there will always be a window between when a transaction has committed on the primary and when it is committed on the secondary, and destruction of the primary in that window will result in lost committed transactions.  The secondary can stay very close to the primary, but if you want zero data loss, you'll seen to use synchronous replication: that's what it is for.
> 
> --
> -- Christophe Pettus
>   xof at thebuild.com
> 
> _______________________________________________
> pgpool-general mailing list
> pgpool-general at pgpool.net
> http://www.pgpool.net/mailman/listinfo/pgpool-general