You notice a high number of waits for the db file scattered read and db file sequential read events
in the recent Automatic Database Diagnostic Monitor (ADDM) report. After further investigation,
you find that queries are performing too many full table scans and indexes are not being used
even though the filter columns are indexed.
Identify three possible reasons for this.
A.
Missing or stale histogram statistics
B.
Undersized shared pool
C.
High clustering factor for the indexes
D.
High value for the DB_FILE_MULTIBLOCK_READ_COUNT parameter
E.
Oversized buffer cache
Explanation:
D: DB_FILE_MULTIBLOCK_READ_COUNT is one of the parameters you can use
to minimize I/O during table scans. It specifies the maximum number of blocks read in one I/O
operation during a sequential scan. The total number of I/Os needed to perform a full table scan
depends on such factors as the size of the table, the multiblock read count, and whether parallel
execution is being utilized for the operation.
ACD
Please check the answer, as C, D, E seems more likely. Oversize buffer cache can prefer full table sacn during CBO decision making. Buffer cache size is one of factor in decision making whether full table sacn would be used or not.
On the same time missing or stale histogram might be related to old stats of index & table, and accordingly to those old stats as well index scan couldn’t be considered in execution path. Only difference would be there that it would go through high db file sequential read but there is no reason it will not hit index with old stats.
A C D
ACD