RAC IMPORTANT QUESTIONS AND ANSWERS

 WHAT IS SCAN CLUSTERWARE AND BENEFITS OF SCAN?

SCAN stands for Single Client Access Name. It's simplifies client connectivity.

 

Single Entry Point: SCAN provides a single, consistent name for clients to access any database instance within the RAC cluster. Now clients only need to know the SCAN name, eliminating the need to manage multiple connection strings.

 

srvctl set scan -scanname myrac-scan.example.com

 

to list the contents of the scan.properties file, you can use the following command:

crsctl list scan

How SCAN Listener Knows the Least-Loaded Node

1. Local Listener Communication: Each node in the RAC cluster has its own local listener that keeps track of the load on its node, including session count, CPU usage, memory usage, and other factors.

• The local listeners regularly communicate with the Global Resource Directory (GRD) in the Oracle Clusterware to update load-related information.

• Each local listener advertises its node’s load information to SCAN listeners.

• If there are 3 voting disks, a node must be able to access at least 2 of them (a simple majority) to continue operating.

• If there are 5 voting disks, at least 3 must be accessible

• Quorum=(3/2)+1=1.5+1=2.5⇒round up to 2

x

WHAT IS VOTING DISK AND ROLE OF VOTING DISK?

VOTING DISK should be created on a shared storage/ASM where all nodes in a cluster register their heartbeat information. This information is used to determine which nodes are active members of the cluster at any given point of time, essentially, it's like an attendance register where nodes mark their presence to confirm they are operational. It is created by oracle clusterware installation process.. In a split-brain situation, the voting disk will determine which node(s) survive and which node(s) will be evicted. Based on the simple majority rule.

What is GRD in Oracle RAC?

GRD Stands for Global Resource Directory and It acts as a central repository for information about Data block locations, Lock information and Resource ownership.

The GRD is distributed across all instances in the cluster, with each instance managing a portion of the directory. Information in the GRD is constantly updated as instances access and modify shared resources.  

The Global Cache Service (GCS) and Global Enqueue Service (GES) processes maintain and update the GRD.

 

Cache fusion ?

Cache fusion is a powerful feature and that leverages on a controlling mechanism called Global Cache Service (GCS), which is responsible for block transfers between instances This significantly reduces disk I/O and improves overall performance. GCS plays a crucial role in coordinating cache fusion by managing block transfers and ensuring data consistency.

The Global Cache Service is implemented by various background processes, such as

Ø  Global Cache Service Processes (LMSn)

Ø  Global Enqueue Service Daemon (LMD)

 

What is GCS service in oracle RAC?

GCS stands for Global Cache Service that ensures data consistency and efficient resource management across multiple instances by Tracking Block Information, Manages Cache Coherency and Facilitates Cache Fusion.

Key Functions of GCS:

Resource Management: Allocates and manages resources like data blocks across instances.  

Cache Coherency: Maintains data consistency by preventing conflicting modifications.  

Performance Optimization: Improves performance through cache fusion and efficient resource utilization.

Fault Tolerance: Contributes to RAC's high availability by managing data block recovery in case of failures.

 

 

Explain the role of the Global Enqueue Service (GES) in oracle RAC?

The Global Enqueue Service (GES) ensures data integrity and consistency. Manages locks on shared resources like library cache, dictionary cache, and transactions. Prevents multiple instances from modifying the same resource simultaneously. When an instance requests a lock, GES checks for conflicts and grants or denies the request accordingly.

Resources Managed by GES:

§  Dictionary Cache Locks: Protects access to data dictionary structures.

§  Library Cache Locks: Protects access to shared SQL and PL/SQL code.

§  Transaction Locks: Ensures data consistency during transactions.

§  DML Locks: Protects data modifications.

§  DDL Locks: Protects schema changes.

 

 

What is a split brain IN ORACLE RAC?

Split brain syndrome occurs in Oracle RAC when the nodes in a cluster lose communication with each other via the private interconnects, but the individual nodes and their database instances are still operational. Both the instances can modify the same data blocks concurrently, leading to inconsistencies and data loss/corruption. This behaviour spilt brain syndrome.

Simple line: The spilt brain syndrome means the effected node in the cluster works independently without commutating with other nodes in the same cluster via interconnect and failed to update heartbeat information in the vote disk. This node will be evicted from the Clusterware to maintain data consistency and integrity.   

 

How it Happens:

Private Interconnect Failure: The high-speed network connecting the RAC nodes goes down.

Isolated Instances: Each node becomes isolated, unaware of the other node's status.  

Independent Operation: Both nodes continue to function, but independently.