Oracle RAC Architecture

What is Oracle RAC?

Oracle RAC allows multiple instances of Oracle Database to run on different servers (nodes) while accessing a single shared database. This architecture not only enhances availability but also provides load balancing and scalability, making it an excellent choice for mission-critical applications.

Key Components of RAC Architecture

1. Nodes
   • Each node in a RAC environment is an independent server running an Oracle Database instance. Multiple nodes work together to process workloads, sharing the database resources efficiently.
2. Shared Storage
   • All nodes access a centralized storage system, usually configured with SAN (Storage Area Network) or NFS (Network File System). This shared storage contains all the essential database files, including data files, control files, and redo log files.
3. Interconnect
   • The interconnect is a high-speed private network that enables communication between the nodes. It is crucial for data transfer, synchronization, and cluster management tasks.
4. Oracle Clusterware
   • This software layer manages the cluster environment. It ensures that instances are operational, coordinates failover processes, and manages resource allocation across the nodes.

Oracle RAC Background Processes

In an Oracle Real Application Clusters (RAC) environment, several background processes work together to ensure efficient management of resources, transaction consistency, and overall database performance

ACMS (Atomic Controlfile to Memory Service)

• The ACMS process is an agent that ensures a distributed System Global Area (SGA) memory update is either globally committed upon success or globally aborted in case of failure. This ensures the integrity and consistency of the control files across the cluster.

GTX0-j (Global Transaction Process)

• The GTX0-j processes provide transparent support for XA global transactions within an Oracle RAC environment. The database autotunes the number of these processes based on the workload associated with XA global transactions, allowing for efficient management of distributed transactions.

LMON (Global Enqueue Service Monitor)

• LMON monitors global enqueues and resources across the cluster. It performs global enqueue recovery operations, ensuring that resource locks are appropriately managed and that the cluster remains in a consistent state.

LMD (Global Enqueue Service Daemon)

• The LMD process manages incoming remote resource requests for each instance. It coordinates access to shared resources, facilitating communication between instances in the RAC.

LMS (Global Cache Service Process)

• The LMS process maintains records of data file statuses and each cached block by tracking information in the Global Resource Directory (GRD). It controls the flow of messages between remote instances and manages global data block access, transmitting block images between buffer caches as part of the Cache Fusion feature.

GES (Global Enqueue Service)

• GES manages global locks and resources in the RAC environment, facilitating communication between instances and ensuring that they can safely access shared data without conflicts.

GCS (Global Cache Service)

• GCS provides cache coherency between instances, allowing them to access shared data blocks consistently. This is crucial for maintaining data integrity across the cluster.

LCK0 (Instance Enqueue Process)

• The LCK0 process manages non-Cache Fusion resource requests, such as those related to library and row caches. It ensures that these resource requests do not conflict with Cache Fusion operations.

RMSn (Oracle RAC Management Processes)

• The RMSn processes perform manageability tasks for Oracle RAC, such as creating resources when new instances are added to the cluster. This ensures smooth scaling and management of the RAC environment.

RSMN (Remote Slave Monitor)

• The RSMN process manages the creation and communication of background slave processes on remote instances. These background slaves perform tasks on behalf of a coordinating process running in another instance, ensuring efficient distributed processing.