After an upgrade to 7.4.1 (from 7.3) we see a severe performance
regression in bulk INSERTs.
This is apparently caused by constant checkpointing (every 10 to 20
seconds). I've already increased the number of checkpoint segments to
32, but currently, there are just 10 or 11 files in the pg_xlog
directory. With 7.3, we had configured checkpoint_segements at 16, and
there were 33 pg_xlog files. Checkpoints happened every couple of
minutes.
How can I reduce the checkpoint frequency?
(I'd like to try that first because it's the most obvious anomaly.
Maybe we can look at the involved table later.)
--
Current mail filters: many dial-up/DSL/cable modem hosts, and the
following domains: atlas.cz, bigpond.com, freenet.de, hotmail.com,
libero.it, netscape.net, postino.it, tiscali.co.uk, tiscali.cz,
tiscali.it, voila.fr, wanadoo.fr, yahoo.com.