Simon Riggs <simon at 2ndquadrant.com> writes:
> On Tue, 2007-11-13 at 10:29 +0100, Cyril SCETBON wrote:
>
>> I think it would be better to be able to modify SLON_DATA_FETCH_SIZE by 
>> using a parameter on the slon command line. For example, we got a lot of 
>> updates (400 per second) between far geographical sites. I had to modify 
>> this variable and to recompile the slon binary cause I didn't find 
>> another way until now.
>
> Cyril,
>
> Could you publish your performance results? That will help everybody to
> understand the benefit of this change.
>
> Do you have results for a variety of settings? It may be that we just
> make a permanent change of the #DEFINEd value, rather than making it
> configurable.

One of our guys is planning to try changing this value, increasing by
an order of magnitude.

I'm inclined to expect that the Right Answer may be to make a
permanent increase to the #DEFINEd value, as you suggest.

 -> It's not easily changed via configuration because the #DEFINEd
    value is used to define, at compile time, an array's size.

 -> It's not a huge array; the elements aren't particularly huge,
    except in the specific case where a tuple is Really Big.

I wouldn't be surprised if shifting from a size of 100 to a size of
1000 would be helpful, and not overly painful in any fashion.
-- 
select 'cbbrowne' || '@' || 'acm.org';
http://www3.sympatico.ca/cbbrowne/oses.html
A Plateau is the highest form of flattery.