What is the expected performance of cluster and what tuning parameters
have most effect?
I have a set of 5 tables with identical structure (all inherit a common
table). The sizes given are total relation size. The clustering index is
a gist index on a (non null) geographic(linestring) column
1. 327600 rows, 105MB, 15.8s
2. 770165 rows, 232MB, 59.5s
3. 1437041 rows, 424MB, 140s
4. 3980922 rows, 1167MB, 276s
5. 31843368 rows, 9709MB, ~ 10 hours
Server is version 9.1. with postgis 1.5.4.
Regards,
Mark Thornton