Hervé Piedvache wrote:
Le mardi 13 Juillet 2004 19:10, Josh Berkus a écrit :
What can I do to get better results ?? (configuration option, and/or
hardware update ?)
What can I give you to get more important informations to help me ?
1) What PostgreSQL version are you using?
2) What's your VACUUM, ANALYZE, VACUUM FULL, REINDEX schedule?
VACUUM FULL VERBOSE ANALYZE;
Every day after the calculation I was talking about ...
3) Can you list the non-default settings in your PostgreSQL.conf?
Particularly, shared_buffers, sort_mem, checkpoint_segments,
estimated_cache, and max_fsm_pages?
sort_mem = 512000
This is too much, you are instructing Postgres to use 512MB
for each backend ( some time each backend can use this quantity
more then one )
vacuum_mem = 409600
max_fsm_pages = 50000000
> max_fsm_relations = 2000
50 milions ? HUG.
what tell you postgres in the log after performing
a vacuum full ?
max_files_per_process = 2000
wal_buffers = 1000
checkpoint_segments = 3
For massive insert you have to increase this number,
pump it up to 16
effective_cache_size = 5000000
5GB for 8 GB system is too much
random_page_cost = 3
on your HW you can decrease it to 2
and also decrease the other cpu costs
BTW, I live in Paris too, if you need a hand...
---------------------------(end of broadcast)---------------------------
TIP 9: the planner will ignore your desire to choose an index scan if your
joining column's datatypes do not match