On 02/15/2010 05:59 PM, Peter Körner wrote: > Peter Körner schrieb: >> River Tarnell schrieb: >>> Peter Körner: >>>> Sure? I stil get the error. I guess it has sth. to do with the ownership >>>> of the spatial_ref_sys table. Is it set to osm? >>> Okay, this is fixed again. I previously fixed it, then it got unfixed when >>> I recreated the table for TS-427. >> >> The query works now. I'll setup the diff import with 1 hour windows >> today. It will then take some days to catch up, I'll then lower the >> window size to 5 minutes and then down to one minute. > > To give you just a rough estimation: it takes ~2 hours to import a 6 > hour window. We are now at 2010/02/02 00:00, so to catch up until today > it will take around 5 day. to catch those 5 days it will take another 2 > days. so all cumulated we'll have an up2date database in 8 or 9 days.
Great, but again it seems rather slow. I have no direct experience with dealing with a full planet ( I don't have powerful enough hardware myself), but from what I have heard from others, it should more be like 1 - 2 hours to apply a daily diff not 8 hours. Especially not with such a fast db server as ptolemy is. So it looks like there might be some oportunity for tuning? May it be worth moving osm2pgsql over to run on the same host to see if that helps in any way? It would perhaps be good to see what the performance bottle neck is though if there is any way to find out. Kai > > Peter > > _______________________________________________ > Maps-l mailing list > [email protected] > https://lists.wikimedia.org/mailman/listinfo/maps-l _______________________________________________ Maps-l mailing list [email protected] https://lists.wikimedia.org/mailman/listinfo/maps-l
