Professional Documents
Culture Documents
Oracle Exadata Overview
Oracle Exadata Overview
Presented By
Joel Goodman
18/09/2009
General Overview
Exadata Benefits
Offload Processing
Hardware Overview
Software Overview
Configuration Overview
Security Realms
Metrics
Resource Management
Migration
Database Servers
Storage Arrays
SQL Offload
I/O Resource Manager
Processing
Storage
Consolidation
(Transparent to
database applications)
Predicate filtering:
Only the rows requested are returned to the database server,
rather than all rows in a table.
Column projection filtering:
Only the columns requested are returned to the database
server, rather than all columns in a table.
Join processing:
Simple star join processing is performed within Exadata.
Backups:
I/O for incremental backups is much more efficient because
only changed blocks are read.
Create/extend tablespace:
Cell formats tablespace extents
SELECT customer_id 1 6
FROM orders Row returned
WHERE order_amount>20000;
Database
Servers
Disk Disk
Exadata Realm
Exadata
Building block of massively parallel Storage Server
Exadata Storage Grid
Up to 1 GB/sec data bandwidth per cell
HP DL180 G5
Does not support nondatabase storage
Software preinstalled
Oracle Exadata Storage Server software
Oracle Enterprise Linux 5.1
Oracle InfiniBand protocol Racked
Exadata
HP Management software Storage
Hardware warranty Servers
3-yr parts/3-yr labor/3-yr on-site
24X7, 4-hour response
Absolutely no custom hardware
12 x 3.5 Disk
Redundant Drives
(SAS or SATA)
110/220V Power
Supplies
2 Intel Xeon
Quad-core
Processors
LO100c
Management Card
Dual-port 4x
InfiniBand DDR
card
16 GB/sec
Scale to 18 Add racks to scale further.
cells in one rack
8 GB/sec
4 GB/sec
Exadata bandwidth
scales linearly with
capacity.
Choice of either
450 GB 15,000 RPM SAS disks (Sports car)
1 TB 7,200 RPM SATA disks (SUV)
Choose SAS Based Servers for High Performance
SAS Advantages SAS SATA Advantage
Throughput (MB/s) 1,000 750 1.33X
Average Seek Time (ms) 3.6 7.4 2.05X
Disk level read errors (per year) 6.3 63 10.00X
Years to disk failure 15.2 11.4 1.33X
iDBProtocol over
Multipath
Storage Switch/Network Aggregated
InfiniBand
iDBProtocol over
Multipath
Storage Switch/Network Aggregated
InfiniBand
/opt/oracle/cell cellcli
/cellsrv/deploy/
config
cellsrv MS
CELLSRV /etc/oracle/cell/network-config
rs ADR
cellinit.ora cellip.ora cellinit.ora
bond0
CELLSRV internal
parameters and
local interfaces IPs
cellcli
cellsrv MS cellcli
cellsrv MS cellcli
cellsrv MS adrci
adrci ssh
adrci ssh/cellcli
CELLSRV
rs CELLSRV
ADR
rs CELLSRV
ADR
eth0 rs ADR eth0
eth0
eth0
Network fabric
1 Initial Network
Preparation
Configuring Hardware
6 ASM Disk Group Configuration
for Exadata of New Exadata
2 Servers
Configuring
5 ASM and Database
3 Configuring
instances
New Exadata Cells
for Exadata
4
Configuring Hosts
to Use Exadata Cells
CellCLI> calibrate
Calibration will take few minutes
Aggregate random read throughput across all luns: 876.62 MBPS
Aggregate read Ios per seconds (IOPS) across all luns: 2820
Controller read throughput: 991.02 MBPS
Lun 6_1 on drive [1I:6_1:1 ] random read throughput: 98.91 MBPS, and 334 IOPS
Lun 6_2 on drive [1I:6_1:2 ] random read throughput: 98.64 MBPS, and 316 IOPS
Lun 6_3 on drive [1I:6_1:3 ] random read throughput: 98.98 MBPS, and 331 IOPS
Lun 6_4 on drive [1I:6_1:4 ] random read throughput: 99.05 MBPS, and 327 IOPS
Lun 6_5 on drive [1I:6_1:5 ] random read throughput: 99.12 MBPS, and 325 IOPS
Lun 6_12 on drive [1I:6_1:12] random read throughput: 98.51 MBPS, and 328 IOPS
CALIBRATE results are within an acceptable range.
CellCLI>
$ cd /etc/oracle/cell/network-config
$ cat /etc/oracle/cell/network-config/cellinit.ora
ipaddress1=192.168.50.23/24
$ cat /etc/oracle/cell/network-config/cellip.ora
cell="192.168.50.27:5042"
cell="192.168.50.28:5042"
cell="192.168.50.29:5042"
Realm
Grid
disk
Exadata Cell
Non-RAC RAC
DB DB
Database-Scoped
Instance Instances
Security Mode
Another ASM cluster
Exadata
Cell
ASM cluster
hosts
A A
CREATE KEY S S
M
/etc/oracle/cell/network.config
M
cellkey.ora
ASSIGN KEY
FOR ASM cluster Each
database
D
B $ORACLE_HOME/admin/<db name>/pfile
cellkey.ora
CREATE/ALTER
Each
disk
GRIDDISK
Each
availableTo ASM disk
D
B
CREATE/ALTER
ASSIGN KEY
CREATE KEY GRIDDISK
FOR DB
availableTo DB
ms keeps a
subset of the metric values
Collected metrics: One hour of LIST METRICCURRENT
Cell, Cell Disks,
Grid Disk, IORM, In-memory
Interconnect metric values
IORM
metric
cellsrv
thresholds
internal
exceeded
errors
Every hour ms
cellsrv
flushes metric
collects cellsrv ADR ms values to disk
metrics
Email
Cell 1 hour
and/or LIST METRICHISTORY
SNMP
Cell LIST ALERTHISTORY
hardware
issues Metric
History
Admin
Disk Disk
Exadata appliances
Intra-database Plan
Inter-database Plan Category Plan
(CELLCLI) (CELLCLI)
Inter-database Plan
Single Inst DB: 34% OLTP : 70%
RAC DB : 66% BATCH : 30%
Category Plan
I/O
Resource
Manager
Across
Context
Dueling Cooperative
Mixed Workload
databases databases
BEGIN
DBMS_RESOURCE_MANAGER.CREATE_SIMPLE_PLAN(SIMPLE_PLAN => 'my_simple_plan',
CONSUMER_GROUP1 => 'high_priority', GROUP1_PERCENT => 80,
CONSUMER_GROUP2 => 'low_priority' , GROUP2_PERCENT => 20);
END;
/
DBA_RSRC_CONSUMER_GROUPS
CONSUMER_GROUP CATEGORY
---------------------------- --------------- dbms_resource_manager.create_category
SYS_GROUP ADMINISTRATIVE
BATCH_GROUP BATCH
INTERACTIVE_GROUP INTERACTIVE
ORA$ MAINTENANCE
OTHER_GROUPS OTHER Category Level 1 Level 2 Level 3
DEFAULT_CONSUMER_GROUP OTHER
LOW_GROUP OTHER Interactive 90%
AUTO_TASK_CONSUMER_GROUP OTHER
Batch 80%
Intradatabase
22%
30%
40%
25%
45%
18%
20%
10%
Interdatabase
70%
30%
70%
30%
Categories
40% 60%
BATCH INTERACTIVE
100%
Source DB Logical
extents size No migration
4MB?
can change
database extent size
Yes
ASM Redundancy
Physical Standby Database
Recovery Manager
Transportable Database
Transportable Tablespace
General Overview
Exadata Benefits
Offload Processing
Hardware Overview
Software Overview
Configuration Overview
Security Realms
Metrics
Resource Management
Migration