[clug] Postgres sync / duplication

Michael Carden crash at michaelcarden.net
Sat May 6 01:52:11 GMT 2006

Thanks for the links Mike J.

On Fri, 5 May 2006 09:07 pm, Martijn van Oosterhout wrote:

> This is probably the wrong list to ask, it'd be better directed at
> pgsql-general (at) postgresql.org. 

Agreed. It's just I'm aware there's some Postgres expertise 'round these 

> In any case, what you're referring 
> to is multi-master replication 

It is, but I'd gathered that it's generally referred to that in circumstances 
where both 'masters' are live to clients all the time. Of course, I'm very 
likely wrong about that.

> and is quite a difficult (and usually 
> expensive) problem in general.

Expensive in what terms? The dollar sort or the system resources sort?

> If you make all changes on the one database, the other can be a
> replicated copy. However, I don't see how you can hope to keep the
> databases in sync without a link...

My idea is that a link would be put in place for the purposes of synch-ing and 
then removed.

I had hoped for (but admittedly not found in the docs) some inbuilt Postgres 
function that lets one server synch itself with another. Because the two live 
on different OSes, I'd hoped to avoid doing it at the filesystem level 
because I just *know* I'll miss something.

> Have a nice day,

Thanks for your comments.


More information about the linux mailing list