Slow performance of query

Question

Hi,
We have been experiencing performance problems in our database lately. Here is the sample query:
select /*+ ORDERED FIRST_ROWS(6) */ * from
(select DCDATA_ID , score(1) as score from aap.DCDATA_imagearc a
where contains(DCDATA_META, '((BASEBALL and (USA not BROWNLOW)) and ((((({aapdefault} within
document@schema) not ((((((((((Credit\~AP or Credit\~AFP) or Credit\~EPA) or Credit\~30) or
Credit\~113) or Cr
edit\~95) or Credit\~13) or Credit\~33) or Credit\~65) or Credit\~56) or Credit\~42)) not
((({ap+arroyo} within credit) or ({ap+chris} within credit)) or ({ap+graylock} within credit))) not
INTL+OUT) not aus
tralia+only))', 1) > 0
order by DCDATA_ID desc )
where ROWNUM <= 6
============
Plan Table
============
------------------------------------------------------------------+---------------------------------
| Id | Operation | Name | Rows | Bytes | Cost | Time
| Pstart| Pstop |
------------------------------------------------------------------+---------------------------------
| 0 | SELECT STATEMENT | | | | 11K |
| | |
| 1 | COUNT STOPKEY | | | | |
| | |
| 2 | VIEW | | 7 | 182 | 11K |
00:02:16 | | |
| 3 | TABLE ACCESS BY GLOBAL INDEX ROWID | DCDATA_IMAGEARC | 14K | 1583K | 11K |
00:02:16 | ROW LOCATION| ROW LOCATION|
| 4 | INDEX FULL SCAN DESCENDING | DCDATA_IMAGEARC_ID| 960 | | 20 |
00:00:01 | | |
------------------------------------------------------------------+---------------------------------
This query waits on "db file sequential read" wait event and takes 5-10 mins and this in turn causes our website to hang. I tried increasing
the SGA_TARGET size from 8GB to 10GB. But, it was of no help. The i gathered statistics but, even
that didn't help. Then i changed the hint FIRST_ROWS in the query to ALL_ROWS. It is definitely working fine
now but, im not convinced that it will be a permanent solution for this.
So, I was thinking of increasing the memory and KEEP pool size and pin the indexes used in the
query into KEEP POOL.
Can you please tell me your views on pinning indexes into KEEP pool.
Is rebuilding index going to help? (We are optimizing the text index daily and I feel rebulding index may not help much).
Appreciate your comments on this.
Thanks,
Ramya

Answer 1

"db file sequential read" means the database is waiting for indexes to be read.
It might be the case that you're reading a very large index or a fragmented one.
Maybe it is smart to rebuild the indexes of the table during the night of even rebuild the table by moving it across two tablespaces