[Tile-serving] [osm2pgsql-dev/osm2pgsql] Logging falsely suggests CLUSTER is run (#1565)

[mboeringa]

Just for the record:

With current PG16 and Facebook Daylights v1.33 **1.1B** records building dataset, CLUSTER using the geometry spatial index is incredibly slow. As far as I can tell from all tests I ran up to now, the performance of CLUSTER in PostgreSQL _deteriorates at more than quadratic rate_ with increasing size of the table.

E.g. with the +/- 500M buildings of the ordinary Planet file, it took about **7h** to spatially defragment the table and rebuild all indexes using CLUSTER, whereas with the 1.1B record dataset, it took **>48 hours**.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/osm2pgsql-dev/osm2pgsql/issues/1565#issuecomment-1802054703
You are receiving this because you are subscribed to this thread.

Message ID: <osm2pgsql-dev/osm2pgsql/issues/1565/1802054703 at github.com>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstreetmap.org/pipermail/tile-serving/attachments/20231108/275200d0/attachment.htm>