--- Tom Lane <tgl@sss.pgh.pa.us> wrote:
> That's pretty difficult to credit after looking at
> the pgstat.c code:
> every incoming blocks_fetched count is added to both
> per-table and
> per-database stats. I wonder if you are looking at
> the wrong per-table
> entries?
i am 100% sure....
"SELECT pg_stat_get_db_blocks_fetched(764755937),
pg_stat_get_db_blocks_hit(764755937);"
gives be constantly increasing stats and
"SELECT relname,heap_blks_read from
pg_statio_user_tables order by heap_blks_read DESC
LIMIT 15;"
is still showing me all zero 4 mins into the test
until i first vacuum analyze is done....
just think abt this.... if we get these stats how
easily we can decide the division of tables in
tablespaces.... just write a simple program which will
collect the data every t mins... analyze it and move
them to diff tablespaces...
is there any other way of finding table usage???
thx a lot tom
Himanshu
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com