Databases Reference
In-Depth Information
| Id | Operation | Name | Rows | Bytes |
---------------------------------------------------------------------------
| 0 | SELECT STATEMENT | | 1 | 8 |
| 1 | TABLE ACCESS BY INDEX ROWID| RCV_TRANSACTIONS | 1 | 8 |
|* 2 | INDEX UNIQUE SCAN | RCV_TRANSACTIONS_U1 | 1 | |
Statistics
----------------------------------------------------------
.. Output omitted
4 consistent gets
4 physical reads
... Output omitted
As you can see, our query used an index called
RCV_TRANSACTIONS_U1
followed by a
TABLE ACCESS BY ROWID
operation to access the
RCV_TRANSACTIONS
table. We then determined the
DATA_OBJECT_ID
for the index and table and
ran the following CellCLI command to display
flashcachecontent
measurements for these objects:
SQL> select data_object_id dobject_id,object_name,owner
2 from dba_objects
3 where owner='PO' and object_name in ('RCV_TRANSACTIONS','RCV_TRANSACTIONS_U1');
DataObjId Object Owner
--------- ------------------------------ --------------
3010199 RCV_TRANSACTIONS PO
3080102 RCV_TRANSACTIONS_U1 PO
[oracle@cm01dbm01 ]$ dcli -g ~/cell_group cellcli -e list flashcachecontent attributes
dbUniqueName,hitCount,missCount,cachedSize,objectNumber where objectNumber=3080102
[oracle@cm01dbm01 ]$ dcli -g ~/cell_group cellcli -e list flashcachecontent attributes
dbUniqueName,hitCount,missCount,cachedSize,objectNumber where objectNumber=3010199
[oracle@cm01dbm01 ]$
As you can see, the
cellcli
command showed no evidence of our segments existing in flash cache. Prior to
running the tests, we bounced our database and recreated our SFC storage in order to conduct a “clean” test; you
would not do this in a production environment, of course, but, here, it is helpful to demonstrate the goal of this recipe.
■
Note
to learn how to drop and create Smart Flash Cache storage, please refer to recipe 18-1.
Prior to executing our previous test query, we ran the script in Listing 18-2 to measure a handful of
relevant statistics.
Search WWH ::
Custom Search