Thanks, I knew about that one, and spent half a day cleaning up the data set before transfer.
That was exciting...
Cheers, Steve
On Thu, 10 May 2007 08:57:25 -0600
Dan Harris <fbsd@drivefaster.net> wrote:
> Steve Holdoway wrote:
> > Are there any gotchas? I've got the opportunity to move to another database server for this application and yould
liketo take the opportunity to upgrade at the same time.
>
> Yes. One of them that got me was that in 8.0.x, you could insert "invalid"
> unicode byte sequences and it would warn you but accept them just fine. In 8.2,
> this is strictly enforced and you will not be able to restore your dump file
> without running a utility to "fix" the dump file.
>
> This change has prevented me from doing the upgrade as my dump file is nearly
> 100GB and the programming design of this utility requires that the whole file be
> read into RAM before it begins processing it ( I've let it run for days and it
> doesn't complete ). I didn't think I had any Unicode in my database at all, but
> I guess there are a few instances of "noise" in there that cause this..
>
> Hopefully this doesn't block someone else from upgrading.
>
> -Dan
>
> ---------------------------(end of broadcast)---------------------------
> TIP 4: Have you searched our list archives?
>
> http://archives.postgresql.org