Andrew Sullivan wrote:
> On Wed, Dec 07, 2005 at 07:09:15PM +1100, Gavin Sherry wrote:
> 
>>There is another problem, however. Slony is not handling databases encoded
>>in different character sets. Say i wanted to upgrade my 8.0 system, which
>>is encoded latin1, to an 8.1 system encoded in utf-8. This is not
>>possible. You will see the same error as above. The thing is, PostgreSQL
>>can do character conversion between different character sets and provides
>>convenient mechanisms, like 'set client_encoding ...', to deal with this.
> 
> 
> Well, I can see an argument either way for this.  One approach might
> be to insist that the two ends be in the same encoding, because
> that's what "replicate" means.  I'm sympathetic to the view, however,
> that we need a knob to control this.

slony is a nice tool to upgrade e.g. from 7.x to 8.x. Manual selection 
would allow encoding cleanup at the same time, so insisting on the same 
encoding would sacrifice reasonable use-cases.

Regards,
Andreas