Sorry. I just figured out it was an endianess problem rather than a header
size problem. Works fine now. Looks like 16 is still the magic number.
Please disregard the last question.
-Tony
"G. Anthony Reina" wrote:
> Thomas,
>
> I've re-done my database with the 32K tuple limit-- looks good.
> However, I seem to be having trouble with binary cursors. I think it may
> be with the number of bytes in the tuple header (used to be 16 bytes
> with the 8K limit). I've tried 16, 32, and 64, but haven't seemed to
> find it. Have you used binary cursors with this setup?
>
> Thanks.
> -Tony
>
> Thomas Swan wrote:
>
> > At 12:24 PM 8/8/2000, G. Anthony Reina wrote:
> >
> >> I know that I've seen this answer before but can't seem to find it
> >> for
> >> 7.0.2 in the archives. Which file(s) need to be changed to have
> >> Postgres
> >> default to 32K size row limits rather than 8K? Has anyone run into
> >> any
> >> horror stories after going to 32K?
> >
> >
> > I've been running it for a while and fairly heavily without any
> > problems...
> >
> > in src/include/config.h modify the following section AFTER running
> > configure.
> >
> > /*
> > * Size of a disk block --- currently, this limits the size of a
> > tuple.
> > * You can set it bigger if you need bigger tuples.
> > */
> > /* currently must be <= 32k bjm */
> > #define BLCKSZ 8192
> >
> > change to
> >
> > #define BLCKSZ 32768
> >
> > This has worked for me....
> > -
> > - Thomas Swan
> > - Graduate Student - Computer Science
> > - The University of Mississippi
> > -
> > - "People can be categorized into two fundamental
> > - groups, those that divide people into two groups
> > - and those that don't."