Quarter Rack of X2-2
Half Rack of X2-2
Full Rack of X2-2
Full Rack of X2-8
Results of random reads.
Results of table scans.
Write to a mirror
All data is cached In the Flash Cache
Load compressed files with Direct Path loading.
Sort incoming data on a column with a low cardinality.
Use DBFS and external flies to load the data.
Use Data Pump to load data.
You can manage workloads within a database with Database Resource Manager.
You can manage workloads across multiple databases with Database Resource Manager
You can manage workloads within a database with I/O Resource Manager
You cannot manage workloads across multiple databases
You use RMAN.
You back up the software files using an operating system method
You do a complete copy of their system storage area.
You do not need to backup this software, as there is an automated recovery mechanism built into each cell.
A grid disk must use all of a cell disk.
A grid disk can only use portions of a cell disk across all cells.
A grid disk can use portions of a cell disk on a subset of cells.
A grid disk can span multiple Exadata Storage Server cells.
The role specified for the category
The role specified for the user
The role specified In a Data Guard environment
The role specified for the application
The cell must be rebooted.
You must alter ASM to alert IT
You must recreate any grid disks that use the drive
Exadata Smart Flash Cache expands the allocation to hold the objects
Not all objects will be stored in Exadata Smart Flash Cache
Overall performance is increased as more objects fit into Exadata Smart Flash Cache
CELL_FLASH_CACHE attribute on the data object
CACHE hint In the SQL statement
Data from a table scan
Small data less than 128 KBs
Control file I/Os
Create a grid disk for every table space
Create two grid disks, one for data and one for recovery
Create two grid disks, one for tables and one for Indexes
Create a grid disk for each table
Based on the priority of the user
Based on the resource group of the user
Based on the intradatabase resource plan
Offload processing allows PL/SQL functions to be executed on the Exadata Storage Server
Offload processing aggregates SQL processing into higher level functions, which are shipped to the Exadata Storage Server
Offload processing moves some processing closer to the stored data, which allows the Exadata Storage Server to return only the rows and columns requested
Offload processing is just marketing - it provides no real advantage
Through flash cache.
Through the Infiniband switch
Through onboard memory
There is never any cell-to-cell communication in an Exadata Storage Server
A Quarter Rack, with 20 TB storage
A Half Rack, with 50 TB raw storage
A Full Rack, with 100 TB of raw storage
You should try to talk your customer into using High Capacity disk
The Exadata Storage Server can select rows based on column values listed in a SQL predicate
Storage indexes are built on columns for use in filtering
Only necessary columns are returned to the database server
Column filtering is a marketing term, not a real benefit
Hardware in the database servers
Software in the database servers
Software in the Exadata Storage Server
Hardware in the Exadata Storage Server
Stores minimum and maximum values for a 1 MB region
Can be used to eliminate cells from queries
Is a standard database index
Persists through a cell reboot
Production over Test and OLTP over maintenance
OLTP over maintenance and control-file I/Os over database writes
Apply operations over read-only queries on standby databases and control-file I/Os over database writes
Apply operations over read-only queries on standby databases and OLTP over maintenance
Use transportable table spaces to migrate data to the new environment
Use Data Guard logical standby to reproduce data on the new system
Use Data Guard physical standby to migrate data to the new system
Use ASM to rebalance data from the old system to the new system
Use Data Pump to move data to the new system
There is no difference - they are different names for the same thing.
The Database Smart Cache can only work on Oracle Exadata Database Machines
The Exadata Flash Cache can only work on Oracle Exadata Database Machines
You can pin tables into the Database Smart Cache
One per Exadata cell
One per Exadata Storage Server
One per DBRM resource group
One per IORM category
Set the IORM parameter to TRUE
Manually set the RESOURCE_MANAGER_PLAN parameter with the ALTER SYSTEM statement
Activate IORMPLAN on each of the target Exadata cells
Activate IORMPLAN for ASM
Storage Indexes and Predicate Filtering
Database Resource Manager
All the time
When there are IORM resource plans in place
When there are more than one resource group for a cell
When I/O requests start to saturate the cell
When I/O requests start to saturate the cell and there is more than one resource group defined for the cell
You only have to drop and add the disk.
You have to drop and add the disk and then rebalance data onto it.
Fast resynch on the disk happens automatically.
Tell the customer that when they need more storage, the underlying hardware will be faster so they will not have to worry.
Tell the customer they can always move from SATA storage to SAS storage for better performance.
Tell the customer that some performance degradation is inevitable in any architecture.
Tell the customer that the balanced hardware configuration of the Exadata Storage Server prevents this from happening.
12 Flash-based cell disks, 12 disk-based cell disks
16 Flash-based cell disks, 16 disk-based cell disks
16 Flash-based cell disks, 12 disk-based cell disks
12 Flash-based cell disks, 16 disk-based cell disks
No Flash-based cell disks, 12 disk-based cell disks
Updates to the Exadata Server software
Executing distributed CU commands
Interaction with the EM Exadata plug-in
Based on Database Manager Resource Groups
Based on an Intra-database Plan
Based on a Category Plan
Based on the priority of the background process
Yes, but only If you use Hybrid Columnar Compression
Yes, but tables or partitions that use Hybrid Columnar Compression will not be immediately accessible
Yes, but only if you do not allow Smart Scans with the Oracle Exadata Database Machine.
You don’t - this customer is not appropriate for the Oracle Exadata Database Machine. Move on
You explain that offload processing works with the same SQL in packaged applications as it does in ad hoc queries
You explain that they can use stored plans to for packaged applications to modify the SQL so that it can perform well with the Oracle Exadata Database Machine.
You try to get them to understand the benefits provided for data warehousing on the same data.
You will add Indexes to your tables for better performance
You should consider implementing a different partitioning strategy.
You should consider reducing the number of materialized views.
You should make the extents larger, such as 8 MBs.
Suggest that this is a great time to upgrade all their databases to RAC, since an Oracle Exadata Database Machine has to have either all RAC or no RAC databases.
Divide the Oracle Exadata Database Machine up into two equal partitions, using one for RAC and the other for their single instance databases.
You can assign databases to database servers in any way, regardless of whether the database is RAC or single instance
Migrate all databases to single instance databases since they always provide better performance on a Oracle Exadata Database Machine
A command representing the SQL statement
The filtering predicates from the current SQL statement
Statistics on data usage
ALTER CELL, CREATE CELLDISK, CREATE GRIDDISK, CALIBRATE
ALTER CELL, CREATE GRIDDISK, CREATE CELLDISK, CALIBRATE
CREATE GRIDDISK, CREATE CELLDISK, CALIBRATE, ALTER CELL
CALIBRATE, CREATE GRIDDISK, CREATE CELLDISK, ALTER CELL
CALIBRATE, ALTER CELL, CREATE GRIDDISK, CREATE CELLDISK
None - they are created by default
Issue the CREATE DISK... FLASHDISK command after initial setup of the cell
Issue DROP FLASHCACHE and then CREATE GRIDDISK FLASHDISK commands
Issue the DROP FLASHCACHE and CREATE FLASHCELLDISK commands
Conflicting workloads, both of which are latency sensitive.
Conflicting workloads, both of which are throughput sensitive.
When I/O is a bottleneck.
When CPU is a bottleneck.
EM Grid Control
EM Database Control
All partitions must either be compressed or uncompressed
All partitions must use either OLTP compression or Hybrid Columnar Compression.
Any partition can use any type of compression, regardless of the compression of other partitions.
You cannot mix compression types among partitions of the same table.
Database names, levels, and allocations
Database names, levels, and resource groups
Database names, allocations, and resource groups
Database names, resource groups, and categories
You can infer from the performance of the statement.
Query statistics from the V$ views
The hints used in a SQL statement
With the CELLCLI SMART.SQL statement
Use HCC in archive mode.
Batch write operations to increase the efficiency of HCC.
Avoid HCC at all costs.
Partition table and use HCC on older, less updated partitions.
You define inter-database resource plans with Database Resource Manager.
You define Intra-database resource plans with Database Resource Manager.
You define categories with Database Resource Manager.
You enable a category plan with CellCLI.
Active data is placed on the outer portion of disk
Mirrored data is placed on the outer portion of disk
Active data from all grid disks is on the inside portion of the disk
Active data and mirrored data are placed together, with the first grid disk going on the outside of the disk.
Oracle Enterprise Linux
SATA disks are always better.
SAS disks could work, but SATA provide more raw bandwidth.
You want to make sure your customer has room for growth, even If they don't think they need it now.
When you include the storage space needed for redundant data, logs, rollback and temp space, a full rack of SATA disks provides 100 TBs of storage for user data.
All data on the Exadata Storage Server is compressed.
All backups from the Exadata Storage Server are compressed.
Recovery Manager always uses more threads to perform backups from an Exadata Storage Server.
Offload processing finds only the changed blocks.
ASM writes all changes that occurred while the disk was offline.
Log entries on the disk are synched with log files.
Data values on the disk are synched with data in the Flash Cache.
There is no synch state for grid disks.
Partitioning and Storage Indexes
Storage Indexes and Parallelism
Partitioning and parallelism
Model scoring and Partitioning
When you create the cell disk
When you create the grid disk
When you create a tablespace
To extend the availability of your ASM disk groups.
To reduce the need for high redundancy
To reduce the amount of data collected for a fast resynch
To increase performance of an Exadata Storage Server.