[Bucardo-general] Replicating Tables the Rely on Sequences

[Ben Allen]

Thanks Greg and Jon for your responses.
The odd/even approach seems like it will work alight. What is your  
blog's URL so I can look for the post?

Thanks for your help.

Regards,

Ben Allen


On Oct 3, 2007, at 11:25 AM, Greg Sabino Mullane wrote:

> On Wed, 2007-10-03 at 11:16 -0600, Ben Allen wrote:
>> Is there a more graceful way to use sequences in a "swap" replication
>> then having the sequence on the second database start at a different
>> number then the first?
>
> Yes, although in some replication systems you don't necessarily  
> want the
> sequence to be the sole unique constraint. For the example you gave, a
> unique index on description might be warranted, followed by a  
> customcode
> exception to sort things out. All of this is a blog post I'm going to
> make soon. As far as sequences, the general options are:
>
> 1) Set one small, one large, as you mentioned
> 2) Set one even and one odd (e.g. INCREMENT BY 2) [I like this one the
> best]
> 3) Have the second database call the sequence on the first via dbilink
> or similar.
> 4) Improve Bucardo to handle things like this by changing the  
> sequences
> on the fly :)
>
>
> -- 
> Greg Sabino Mullane greg at endpoint.com
> End Point Corporation 610-983-9073
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: https://mail.endcrypt.com/pipermail/bucardo-general/attachments/20071003/746ff958/attachment.html