Oracle RAC (Real Application Clusters)

Oracle RAC (Real Application Clusters) allows multiple servers (nodes) to run one database, providing high availability, scalability, and fault tolerance. This post covers essential RAC concepts, architecture, components, and troubleshooting tips to help DBAs prepare for interviews.

1. What is Oracle RAC and why is it used?
Oracle RAC lets multiple servers work together to run a single database, providing continuous availability and better performance by spreading workload.

2. Explain the architecture of Oracle RAC.
RAC has multiple nodes connected to shared storage, with Oracle Clusterware managing coordination, and Cache Fusion technology ensuring data consistency across nodes.

3. What are the key components of Oracle RAC?
Clusterware, shared storage, multiple database instances (one per node), Voting Disk, Oracle Cluster Registry (OCR), and Cache Fusion.

4. How does Oracle RAC provide high availability and scalability?
By allowing nodes to failover automatically and distributing user workload across nodes, RAC avoids downtime and scales performance.

5. What is the role of the Clusterware in RAC?
Clusterware manages node membership, resource monitoring, failover, and cluster communication.

6. How do nodes in RAC communicate?
Nodes communicate over a private interconnect using Cache Fusion to share and manage data blocks.

7. What is the Cache Fusion technology in RAC?
Cache Fusion allows instances to share data blocks directly over the interconnect instead of reading from disk, improving performance and data consistency.

8. Explain Global Cache Service (GCS) and Global Enqueue Service (GES).
GCS manages cache coherency for data blocks, and GES manages locks (enqueues) on resources cluster-wide.

9. What are the typical background processes in Oracle RAC?
In addition to normal Oracle processes, RAC has processes like LMON (Lock Manager), LMD (Lock Daemon), LMS (Lock Manager Server), and others for cluster coordination.

10. How is data consistency maintained across RAC nodes?
Through Cache Fusion and Global Cache Services, RAC ensures that data blocks are consistent and properly locked.

11. What is the role of the Voting Disk and OCR (Oracle Cluster Registry)?
Voting Disk helps determine cluster node membership; OCR stores cluster configuration data.

12. How do you add or remove a node from a RAC cluster?
Using Oracle Clusterware commands like addnode.sh or removenode.sh, plus configuring the database instances accordingly.

13. What is the difference between RAC and Data Guard?
RAC is for high availability and scalability within one site (shared storage), while Data Guard provides disaster recovery by maintaining standby databases at separate sites.

14. How is load balancing handled in RAC?
Load balancing is done at the client side using SCAN and services, and internally by distributing sessions across nodes.

15. How does RAC handle failover and node failures?
Clusterware detects failures and automatically fails over sessions to surviving nodes without downtime.

16. What are common RAC performance issues and how to troubleshoot them?
Issues like interconnect latency, cache fusion bottlenecks, and locking conflicts; troubleshoot using AWR, ADDM reports, and monitoring cluster interconnect.

17. How is backup and recovery managed in RAC environments?
Using RMAN with support for RAC, backups can be taken from any node or the shared storage.

18. How do services work in RAC?
Services represent workloads and can be assigned to one or more nodes for better resource management and failover.

19. What is SCAN (Single Client Access Name) in RAC?
SCAN provides a single DNS name to clients for connecting to RAC, simplifying client configuration and load balancing.

20. What are the key initialization parameters specific to RAC?
Parameters like cluster_database, instance_number, thread, service_names, and scan_name help configure RAC behavior.