Oracle 10g rac_overview

32
Oracle RAC Oracle RAC Overview of Real Application Overview of Real Application Clustering Features and Clustering Features and Functionality Functionality

Transcript of Oracle 10g rac_overview

Oracle RACOracle RAC

Overview of Real Application Overview of Real Application Clustering Features and Clustering Features and FunctionalityFunctionality

OverviewOverview

What is RAC?What is RAC? Cache FusionCache Fusion Failover and Load-balancing Failover and Load-balancing Transparent Application Failover Transparent Application Failover

(TAF) (TAF) Other RAC FeaturesOther RAC Features

RAC – What is it?RAC – What is it?

Multiple instances of Oracle running on up Multiple instances of Oracle running on up to 8 nodes to 8 nodes

Multiple instances share a single physical Multiple instances share a single physical databasedatabase

All instances can simultaneously execute All instances can simultaneously execute transactions against the single databasetransactions against the single database

Caches are synchronized using Oracle’s Caches are synchronized using Oracle’s Global Cache Management technology Global Cache Management technology (Cache Fusion)(Cache Fusion)

Previous Oracle Clustering ProductsPrevious Oracle Clustering Products Oracle FailSafe on WindowsOracle FailSafe on Windows OPS (Oracle Parallel Server) on multiple OPS (Oracle Parallel Server) on multiple

platformsplatforms OPS to RAC: OPS to RAC: 7.3 OPS 7.3 OPS 8i OPS 8i OPS 9i RAC 9i RAC

10g RAC10g RAC

The clustering mechanism used to The clustering mechanism used to be more dependent on the be more dependent on the Operating System.Operating System.

With 10g RAC, clustered database is With 10g RAC, clustered database is built into Oraclebuilt into Oracle

History of Oracle RACHistory of Oracle RAC

Oracle RAC FeaturesOracle RAC Features Full Cache FusionFull Cache Fusion Enhanced coordination of cache Enhanced coordination of cache

management and distributed lock management and distributed lock manager (DLM)manager (DLM)

Lock simplification and automationLock simplification and automation Global Cache Service coordinates local Global Cache Service coordinates local

buffer cache and remote block transfersbuffer cache and remote block transfers Enhanced IPCEnhanced IPC Resource configuration simplification and Resource configuration simplification and

automationautomation Improved cluster aware tools Improved cluster aware tools Enhanced DBCAEnhanced DBCA Oracle Enterprise Manager and Grid Oracle Enterprise Manager and Grid

Control Integration EnhancementsControl Integration Enhancements

RAC uses “Shared RAC uses “Shared Everything”Everything”

Users

Database

Server Server Server Server

How RAC clustering is How RAC clustering is donedone One set of dataOne set of data All nodes in the cluster see the same All nodes in the cluster see the same

set of dataset of data All nodes have accessAll nodes have access Any node can update the dataAny node can update the data

Increased ManageabilityIncreased Manageability

One virtual system to configure and One virtual system to configure and managemanage Single Oracle Database Single Oracle Database Single management consoleSingle management console Single system image for the database Single system image for the database

integrated with the clusterintegrated with the cluster Cluster-wide monitoring and diagnostics Cluster-wide monitoring and diagnostics Oracle Enterprise Manager Integration (9i)Oracle Enterprise Manager Integration (9i) Oracle DBConsole and Grid Control Oracle DBConsole and Grid Control

Integration (10g)Integration (10g)

What’s shared; What’s notWhat’s shared; What’s not SharedShared

Disk accessDisk access Resources that manage dataResources that manage data All instances have common data & controls All instances have common data & controls

filesfiles Not SharedNot Shared

Each node has its own dedicated:Each node has its own dedicated: System memorySystem memory Operating systemOperating system Database instanceDatabase instance Application softwareApplication software

Each instance has individual Each instance has individual Log files and Log files and Rollback segmentsRollback segments

RAC can performRAC can perform

Load-balancingLoad-balancing FailoverFailover

Load-Balancing through the Load-Balancing through the ListenerListener

Database

Node 4 Node 1 Node 2 Node 3

Client

Listener

List

ener Listener

Listener

How workload is balancedHow workload is balanced

Client

Node 1

Node 2

Database

Nodes report Nodes report CPU usage to CPU usage to listenerslisteners

How workload is balancedHow workload is balanced

Client

Node 1

Node 2

Database

Listeners choose Listeners choose least busy node least busy node when request when request comes in from comes in from clientclient

Load-BalancingLoad-Balancing

Users

Database

Node 4 Node 1 Node 2 Node 3

FailoverFailover

If a node in the shared disk cluster fails, the If a node in the shared disk cluster fails, the system dynamically redistributes the workload system dynamically redistributes the workload among the surviving cluster nodes.among the surviving cluster nodes.

RAC checks to detect node and network failures. RAC checks to detect node and network failures. A disk-based heartbeat mechanism uses the A disk-based heartbeat mechanism uses the control file to monitor node membership and the control file to monitor node membership and the cluster interconnect is regularly checked to cluster interconnect is regularly checked to determine correct operation. determine correct operation.

Reduced time to recovery with concurrent Reduced time to recovery with concurrent resource configuration and instance (cache) resource configuration and instance (cache) recoveryrecovery

Enhanced failover reliability in 10g with the use Enhanced failover reliability in 10g with the use of Virtual IP addresses (VIPs)of Virtual IP addresses (VIPs)

FailoverFailover

Users

Database

Server Server Server Server X

Transparent Application Transparent Application FailoverFailover Masks failures to end users; they don’t Masks failures to end users; they don’t

need to log back into the systemneed to log back into the system Applications and users are transparently Applications and users are transparently

reconnected to another nodereconnected to another node Applications and queries continue Applications and queries continue

uninterrupteduninterrupted Transactions can failover and replayTransactions can failover and replay

Login context maintainedLogin context maintained DML transactions are rolled backDML transactions are rolled back

RAC Improvements for RAC Improvements for Oracle 9iOracle 9i Full Cache FusionFull Cache Fusion Enhanced coordination of cache Enhanced coordination of cache

management and distributed lock management and distributed lock manager (DLM)manager (DLM)

Lock simplification and automationLock simplification and automation Global Cache Service coordinates local Global Cache Service coordinates local

buffer cache and remote block transfersbuffer cache and remote block transfers Enhanced IPC (InterProcess Enhanced IPC (InterProcess

Communication)Communication) Resource configuration simplification and Resource configuration simplification and

automationautomation

Oracle 10g RAC New Oracle 10g RAC New FeaturesFeatures Integrated Clusterware ManagementIntegrated Clusterware Management

No third-party clusterware software requiredNo third-party clusterware software required Automatic Workload ManagementAutomatic Workload Management

Application workloads can be managed through named Application workloads can be managed through named servicesservices

Single System Image ManagementSingle System Image Management Enterprise Manager manages RAC instances as a single imageEnterprise Manager manages RAC instances as a single image

Fast Connection FailoverFast Connection Failover Fast recovery between the database and mid-tier applicationsFast recovery between the database and mid-tier applications

Performance ImprovementsPerformance Improvements Reduced message traffic, memory usage, and other resourcesReduced message traffic, memory usage, and other resources

Zero Downtime PatchingZero Downtime Patching Patches may be applied one node at a time without downtimePatches may be applied one node at a time without downtime

Cluster Verification and Improved Diagnostic ToolsCluster Verification and Improved Diagnostic Tools New cluster diagnostic tool and improved diagnostic toolsNew cluster diagnostic tool and improved diagnostic tools

Full Cache FusionFull Cache Fusion Is a major feature of RAC starting with 9iIs a major feature of RAC starting with 9i The underlying technology that enables The underlying technology that enables

RACRAC Protocol that allows instances to combine Protocol that allows instances to combine

their data caches into a shared global their data caches into a shared global cachecache

Allows any node to get the most up-to-Allows any node to get the most up-to-date data information from the cache of date data information from the cache of any other node in the cluster without any other node in the cluster without having to access the disk drives again.having to access the disk drives again.

Improved performance with 10gImproved performance with 10g

What is Cache Fusion? What is Cache Fusion? When do I care about it?When do I care about it? ““Dirty” block of data is createdDirty” block of data is created

Data from disk is read into memory on a Data from disk is read into memory on a nodenode

Data is updated on that nodeData is updated on that node Data hasn’t been written to disk yet.Data hasn’t been written to disk yet.

Another node requests the dataAnother node requests the data

““ABC” block of data written ABC” block of data written to the disk drives in the to the disk drives in the databasedatabase

Node A Node B

ABC Data

““ABC” block of data read ABC” block of data read into memory on Node Ainto memory on Node A

Node A Node B

ABC Data

ABC Data

““ABC” updated to “XYZ” in ABC” updated to “XYZ” in cachecache

Node A Node B

ABC Data

ABC DataXYZ Data

Node B requests data Node B requests data blockblock

Node A Node B

ABC Data

ABC DataXYZ Data

I want data!Gimm

e! Gimm

e!

Node A must write data Node A must write data block to disk driveblock to disk drive

Node A Node B

ABC DataXYZ Data

I want data!Gimm

e! Gimm

e!

ABC DataXYZ Data

write

Previous t

o 9i RAC

Node B must read data block Node B must read data block from disk drivefrom disk drive

Node A Node B

ABC DataXYZ Data XYZ Data

ABC DataXYZ Data

read

Previous t

o 9i RAC

Now with RAC Cache Now with RAC Cache FusionFusion

Node A Node B

ABC Data

ABC DataXYZ Data

I want data!Gimm

e! Gimm

e!XYZ Data

Data is transferred Data is transferred immediately via the immediately via the interconnectinterconnect

Shared cache minimizesShared cache minimizesslow I/Oslow I/O

Shared Cache Across Shared Cache Across NodesNodes

Users

Database

Server Server Server Server

Resource Simplification and Resource Simplification and AutomationAutomation

No init.ora parameters requiredNo init.ora parameters required Resource affinityResource affinity

to move the location of the resource masters to move the location of the resource masters for a database file to the instance where block for a database file to the instance where block operations are most frequently occurring. This operations are most frequently occurring. This optimizesoptimizes

Dynamic resource remasteringDynamic resource remastering Ability to move the ownership of a resource Ability to move the ownership of a resource

between instances of Real Application between instances of Real Application Clusters. Clusters.

Dynamic resource remastering is used to Dynamic resource remastering is used to implement resource affinity for increased implement resource affinity for increased performance. performance.

ReviewReview

What does cache fusion avoid that was What does cache fusion avoid that was mandatory in previous versions of Oracle mandatory in previous versions of Oracle Parallel Server?Parallel Server?

Which Oracle process is most important Which Oracle process is most important in managing user session failover?in managing user session failover?

If the purpose of the interconnect is NOT If the purpose of the interconnect is NOT to serve as a “heartbeat”, where is the to serve as a “heartbeat”, where is the heartbeat? heartbeat?

SummarySummary

New FeaturesNew Features Shared Everything ClusteringShared Everything Clustering Cache FusionCache Fusion RAC Clustering failover & load-RAC Clustering failover & load-

balancingbalancing