You may well be interested in point-in-time-recovery aka PITR aka 
continuous backup aka commit log file archiving and shipping. Simpler 
than Slony in your case, and no inside-the-db style changes to your 
schema or runtime.

See http://www.postgresql.org/docs/8.0/interactive/backup.html , 
specifically 
http://www.postgresql.org/docs/8.0/interactive/backup-online.html.

But, in direct answer to your question, running ntpd on slony master 
and replica ought to suffice.


On May 26, 2005, at 10:24 AM, Darrell A. Sullivan, II wrote:

> Hello,
>
> I am considering a few different open source databases and one of my 
> big
> requirements is the ability to replicate my database for backup 
> purposes.
>
> My major goal is not high availability. If the database becomes 
> unavailable
> for several hours it will not be a tragedy. What I do want to make 
> sure of
> is that all of my data updates are replicated to another server in a 
> secured
> location a thousand miles from the main server for disaster recovery
> purposes.
>
> It appears that Slony is the proposed solution for this on PostgreSQL.
>
> I have read that Slony requires the clock on the master and server to 
> be
> synchronized. Can someone tell me why this is and what the 
> consequences are
> if they are out of synch? Also, to what level of granularity must they 
> be in
> Synch? In my experience the time keeping function on PCs is pretty
> miserable. I would not want to have to trust the security of my data 
> to such
> innacurate clocks.
>
> Thanks,
> Darrell
>
> _______________________________________________
> Slony1-general mailing list
> Slony1-general at gborg.postgresql.org
> http://gborg.postgresql.org/mailman/listinfo/slony1-general
>
----
James Robinson
Socialserve.com