How to reduce disk usage and found where disk usage is used? + reindex force doesn`t seem to work

Hi all,

  I'll really appreciate any help to reduce the disk usage of
postgresql.  I have a web site witch is data are refreshed each night.
Right now the disk usage is about 400 Megs but since I reload data all
nights it getting huge.

I do vacuum each time I am finished loading data.


I look into the db to found where space are used .. here is what I got
SELECT relname, (relpages*8)/1024 FROM pg_class ORDER BY relpages DESC;
relname     ?column?
il_idx_1     155
il     83
vd_pk     67
il_pk     61
ic     2
ic_pk     0
iq     0



I see that index on IL (table with 4 column and ~1500000 rows) take
about 155 megs.  Is that normal that IL has 83 megs but the index 155 ?

Vd_pk take about 67 megs but it the table is currently empty.  Is that
means vacuum don`t shrinks indexes?


I tried a reindex with pgadminIII, here is the result.  I noticed that
only system table were reindexed!  Is that normal?

NOTICE:  table "pg_class" was reindexed
NOTICE:  table "pg_attrdef" was reindexed
NOTICE:  table "pg_constraint" was reindexed
NOTICE:  table "pg_description" was reindexed
NOTICE:  table "pg_proc" was reindexed
NOTICE:  table "pg_rewrite" was reindexed
NOTICE:  table "pg_type" was reindexed
NOTICE:  table "pg_attribute" was reindexed
NOTICE:  table "pg_index" was reindexed
NOTICE:  table "pg_operator" was reindexed
NOTICE:  table "pg_opclass" was reindexed
NOTICE:  table "pg_am" was reindexed
NOTICE:  table "pg_amop" was reindexed
NOTICE:  table "pg_amproc" was reindexed
NOTICE:  table "pg_language" was reindexed
NOTICE:  table "pg_aggregate" was reindexed
NOTICE:  table "pg_trigger" was reindexed
NOTICE:  table "pg_cast" was reindexed
NOTICE:  table "pg_namespace" was reindexed
NOTICE:  table "pg_conversion" was reindexed
NOTICE:  table "pg_depend" was reindexed
NOTICE:  table "pg_statistic" was reindexed
NOTICE:  table "pg_inherits" was reindexed
NOTICE:  table "pg_largeobject" was reindexed


I look at directory pg_xlog with is 164 megs.  I change
checkpoint_segments to 1 in postgresql.conf to reduce this.


I have 2 questions:
-Is there any other stuff I can do to reduce disk space?
-What can I do to have à disk usage with is stable?  Remember that I
flush/load data each night.

Thanks for your help
/David






Total query runtime: 141 ms.