[OSM-dev] Potlatch - Richard... I know

[Stefan de Konink]

80n wrote:
> On Tue, Mar 31, 2009 at 5:20 AM, Stefan de Konink <stefan at konink.de 
> <mailto:stefan at konink.de>> wrote:
> 
>     I really wonder, considering the claims of the software that I am using
>     right now, if the API 0.6 upgrade adds referential integrity at 10x the
>     cost I see now, boy, that will be unworkable, considering that this is
>     already running on a 64GB system.
> 
> Stefan
> It isn't clear to me what you are trying to say here. 
> 
> Perhaps I'm missing some context as the rest of your post seems to be 
> about Potlatch. 
> 
> Is there any chance you could clarify this statement for me?  Do you see 
> some problem that we ought to be worried about?

I am running a DBM that in test is roughly 10x faster than PostgreSQL, 
on TPC-H 100. That is basically a dataset of 100GB, the current tables 
of OSM are 'in data' also ~100GB. If the performance that I now see for 
checking constraints using hardware that is much bigger than the server 
that OSM is going to buy/has bought, I start the worry. Especially about 
the time that is reserved for the migration. I think you will not get 
the foreign key constraints checked in a weekend, not even if all 
foreign key constraints are right, which is not the case as you see in 
the OSM Fixer emails.

As I wrote on the Dutch mailinglist, the work I am doing now is 
basically to make sure an API0.6 migration can go smooth at all.


Stefan