03 b sc enablement exadata asm

8/10/2019 03 B SC Enablement Exadata ASM

1/28

Automatic Storage ManagementTechnical Overview


2/28

The DBA Storage Challenges

Databases are rapidly growing in size DBAs have to manage thousands of data files

Decreasing windows of scheduled downtime

Storage resources involve many organizations

Database Admin Systems Admin

Storage Admin

Network Admin Management of database storage is increasing in complexity and

cost


3/28

Storage Management Gap

0

10

20

30

4050

60

1997

1998

1999

2000

2001

2002

2003

2004

2005

Strategic Research Corporation

Capacity in

Terabytes

DiskCapacity

Capacity/DB Adm

Avg. Managed Capacity per Administrator

I n Database Sites with >100 GB Capacity

Mgmt Gap


4/28

Automatic Storage Management

A Database File Systemwhich provides cluster file system and

volume manager capabilities that are integrated into the Oracledatabase 10g kernel at no additional cost

Reduce Cost and Complexity Without CompromisingPerformance or Availability

Simplifies and Automates Database storage Management

Increases Storage Utilization and Agility

Predictably Delivers on Performance and Availability SLAs


5/28

Storage

Arrays

Database Storage

Provisioning and Management

ASM also simplifiesthis process for

clustered storage

This is where ASM

fits into the stack

ServersApplication &DB mgmt

File System Mgmt

Volume Mgmt

Zoning

LUN Masking

Assign Volume Ports

Configure LUNs

Networked

Storage

Networked

storage


6/28

AutomaticStorage

Management

The Operational Stack

Disks

Logical Vol

File System

0010 0010 0010 0010 0010

0010 0010 0010 0010 0010Files

Tablespace

Tables

Disk Group

Logical Vol

File System

File Names

Tablespace

Tables

Before ASM ASM

Network ed Storage

(SAN, NAS, DAS)


7/28

Interaction Between Database

Instances and ASM

Screenreader hint: Thisslidecontainsa diagramthatis

explainedin thespeaker notes.

ASM Instance

System Global Area

Shared

Pool

Large

Pool

ASM

Cache

Processes

RBAL

ARBn

GMON

Onnn

PZ9n

MARK

Database Instance

System Global Area

Shared

Pool

Large

Pool

Buffer

Cache

Processes

PMON

DBWR

SMON

Snnn

CKPT

LGWR

OtherOther

Other

OtherOther

Other

ASM Storage

CreateFi

le

ExtentM

ap

Commit

Close

Initialize

Allocation

1

2

3

4

56


8/28

Disk Group Overview

Disk 1 Disk 2 Disk 3 Disk 4 Disk 5 Disk 6

File 1 File 2

File 3 File 4

Screenreader hint: Thisslidecontainsa diagramthatis

explainedin thespeaker notes.


9/28

ASM Disk Groups

A pool of disks managed as alogical unit

Disk Group


10/28


11/28

ASM Disk Groups

A pool of disks managed as alogical unit

Divides total disk space into

uniform sized megabyte units

ASM spreads each Oracle fileevenly across all disks in a disk

group

Disk Group


12/28

Traditional vs ASM Add Disk

1. Add Disk to OS

2. Create volume(s) withVolume Manager

3. Create File System over

volume4. Figure out data to move to

new disk

5. Move data to new files

6. Rename files in database

7. Re-tune I/O

1. Add Disk to OS

2. Issue the Add Diskcommand


13/28

Adding a Disk with ASM

DBA issues the request to

add a disk to the existing

disk group

Disk Group


14/28

ASM Dynamic Rebalancing

Automatic online rebalancewhenever storage configuration

changes

Disk Group


15/28

ASM Dynamic Rebalancing

Automatic online rebalancewhenever storage configuration

changes

Only move data proportional to

storage added

Disk Group


16/28

Traditional vs ASM Remove Disk

1. List all data that is on disk

2. Choose existing filesystem

to hold data from dropped

disk

3. Move data to new files4. Rename files in database

5. Remove disk from OS

6. Re-tune I/O

1. Issue drop disk command

2. Remove the disk from OS

when Oracle says it is done


17/28

Traditional vs ASM Tune I/O

1. Daily monitor I/O

performance

2. Discover hot spots

3. Figure out how to remove

hot spots4. Move data to new files

5. Rename files in database

6. Verify that hot spot is gone

This space intentionally left

blank


18/28

Data Distribution

Automatic Storage Management Extends SAME

Allows dynamic online Storage reconfiguration

Efficient relocation of data during rebalance Eliminate manual I/O tuning in all storage configurations

Unstriped Disks Striped Disks


19/28

Data Mirroring with ASM

Choices for disk groupredundancy

External: defers to hardware

mirroring

Normal: 2-way mirroring

High: 3-way mirroring

Integration with database

removes need for dirty

region logging


20/28

ASM Mirroring

Mirror at extent level

Mix primary & mirror extents on each disk


21/28

ASM Mirroring

Mirror at extent level

Mix primary & mirror extents on each disk


22/28

ASM Mirroring

No hot spare disk required

Just spare capacity

Failed disk load spread among survivors

Maintains balanced I/O load


23/28

ASM Failure Groups

Failure Group 1 Failure Group 2

Controller 1 Controller 2

Disk Group

Disk Groups assigned by DBAs

AFailure Groupis a set ofdisks sharing a common

resource whose failure needsto be tolerated Redundant copies of an extent

are stored in separate FailureGroups

Failure Groupsassigned byDBAs or automatically by ASM


24/28

ASM Architecture

Pool of Storage

ASM

Instance

Server

NonRAC

Database

Oracle

DB Instance

Disk Group


25/28

ASM Enables Consolidated Clustered

Storage

Clustered

Pool of Storage

ASM

Instance

ASM

Instance

ASM

Instance

ASM

Instance

Clustered

Servers

RAC orNonRAC

Databases

Oracle

DB Instance

Oracle

DB Instance

Oracle

DB Instance

Oracle

DB Instance

Disk GroupDisk Group

Database Database

ASM

Instance

Oracle

DB Instance

Screenreader hint: Thisslide

containsa diagramthatisexplainedin thespeaker notes.


26/28

ASM Dynamic Performance Views

Diagram

V$ASM_DISKGROUP

V$ASM_ATTRIBUTE

V$ASM_DISK

V$ASM_ALIASV$ASM_FILE

V$ASM_TEMPLATE

V$ASM_OPERATION

V$ASM_CLIENTV$ASM_DISK_IOSTAT


27/28

Lower Cost and Complexity

Without Compromising

Performance or Availability

Simplify and Automate

Database Storage

management

Increase Storage Utilizationand Agility

Predictably Deliver on

Service Level Agreements

Fraction of the time is needed tomanage database files

Eliminate over provisioning andmaximize storage resourceutilization with new flexibility

Never get out of tune deliveringhigher performance than raw and

file system sustained over time

Uncompromised availabilityempowering low cost storagedeployment reliably


28/28

Copyright 2009, Oracle Corporation and/or its affiliates

28

03 b sc enablement exadata asm

Documents