10.2.0.1 (Linux) DB will not cache rows from large IOT

Question

Hi I have a 1.6gig IOT, I'll call "mytable" that I'm trying to get the DB to cache rows from. I have done an
alter table mytable storage(buffer_pool keep);
and when I query the IOT index from dba_indexes it says KEEP for buffer_pool.
My db_keep_cache_size is set to 912M, db_cache_size=1G, sga_target=0, sga_max_size=2G.
The problem that I can see is this, the keys for the table are name and day when I say
select * from mytable where name='HIHI' and day='22-FEB-10';
it reads from disk every time. I have turned autotrace on and it is doing a FAST FULL SCAN of the IOT index.
How do I force Oracle to put these rows into the keep cache and read from memory?
I'm using DB - 10.2.0.1
Thanks.

Answer 1

1. If I run that query shown above it give me this -->
TZTEST (RUN FROM DB HOME)
The patch has been applied correctly
I also ran the following -->
SQL> select to_timestamp_tz('2007-03-28 10:00:00 US/EASTERN','YYYY-MM-DD HH24:MI:SS TZR') from dual;
TO_TIMESTAMP_TZ('2007-03-2810:00:00US/EASTERN','YYYY-MM-DDHH24:MI:SSTZR')
28-MAR-07 10.00.00.000000000 AM US/EASTERN
Is there something else I can run to verify the time in sqlplus?
2. The time on the OS is using NTP and it is correct. If I push the date back on the OS to a month ago, the EM time is fine. But with today's date that was affected due to 2007 dst, then it is screwed up.
How can I see the time change from SQLPLus?
Please let me know. Thank you
Message was edited by:
user602765