lonvia left a comment (osm2pgsql-dev/osm2pgsql#37)
Two thoughts on this with regard to the current osm2pgsql implementation:
The only mutation of the tables containing the intarray gin indexes are now
done within a bulk COPY context. Having fastupdate=off may not even be a
performance issue aymore, if PostgreSQL does a bulk update of the index for
COPY, then leaving fastupdate=off in place might do exactly what we'd want.
We used to have clean read and write phases, so that using
`gin_clean_pending_list()` used to make sense after the write phase. Nowadays,
we officially allow reading from the middle tables. The write phase might
coincide with reads from the user. Switching fastupdate on might then be a
performance improvement for the database updates but involuntarily slow down
user actions.
Eventually we need some performance experiments here to see if changing of our
current setup is worth it.
--
Reply to this email directly or view it on GitHub:
https://github.com/osm2pgsql-dev/osm2pgsql/issues/37#issuecomment-3213670771
You are receiving this because you are subscribed to this thread.
Message ID: <osm2pgsql-dev/osm2pgsql/issues/37/[email protected]>
_______________________________________________
Tile-serving mailing list
[email protected]
https://lists.openstreetmap.org/listinfo/tile-serving
