Beginners Guide to Oracle RAC Architecture

Oracle RAC: Overview

A cluster comprises multiple interconnected computers or servers that appear as if they were one server to end users and applications. Oracle RAC enables you to cluster an Oracle database. Oracle RAC uses Oracle Clusterware for the infrastructure to bind multiple servers so they operate as a single system.

Oracle Clusterware is a portable cluster management solution that is integrated with Oracle Database. Oracle Clusterware is also a required component for using Oracle RAC. In addition, Oracle Clusterware enables both noncluster Oracle databases and Oracle RAC databases to use the Oracle high-availability infrastructure. Oracle Clusterware enables you to create a clustered pool of storage to be used by any combination of noncluster and Oracle RAC databases.

Oracle Clusterware is the only clusterware that you need for most platforms on which Oracle RAC operates. You can also use clusterware from other vendors if the clusterware is certified for Oracle RAC. Noncluster Oracle databases have a one-to-one relationship between the Oracle database and the instance. Oracle RAC environments, however, have a one-to-many relationship between the database and instances. An Oracle RAC database can have up to 100 instances, all of which access one database. All database instances must use the same interconnect, which can also be used by Oracle Clusterware. You can also scale up to 64 reader nodes per Hub Node.

Oracle RAC databases differ architecturally from noncluster Oracle databases in that each Oracle RAC database instance also has:

  • At least one additional thread of redo for each instance
  • An instance-specific undo tablespace

The combined processing power of the multiple servers can provide greater throughput and Oracle RAC scalability than is available from a single server.

Typical Oracle RAC Architecture

Oracle RAC architecture

The image above shows how Oracle RAC is the Oracle Database option that provides a single system image for multiple servers to access one Oracle database. In Oracle RAC, each Oracle instance must run on a separate server.

Traditionally, an Oracle RAC environment is located in one data center. However, you can configure Oracle RAC on an extended distance cluster, which is an architecture that provides extremely fast recovery from a site failure and allows for all nodes, at all sites, to actively process transactions as part of a single database cluster. In an extended cluster, the nodes in the cluster are typically dispersed, geographically, such as between two fire cells, between two rooms or buildings, or between two different data centers or cities. For availability reasons, the data must be located at both sites, thus requiring the implementation of disk mirroring technology for storage.

If you choose to implement this architecture, you must assess whether this architecture is a good solution for your business, especially considering distance, latency, and the degree of protection it provides. Oracle RAC on extended clusters provides higher availability than is possible with local Oracle RAC configurations, but an extended cluster may not fulfill all of the disaster-recovery requirements of your organization. A feasible separation provides great protection for some disasters (for example, local power outage or server room flooding) but it cannot provide protection against all types of outages. For comprehensive protection against disasters, including protection against corruptions and regional disasters, Oracle recommends the use of Oracle Data Guard with Oracle RAC.

Oracle RAC One Node

Using Oracle RAC One Node online database relocation, you can relocate the Oracle RAC One Node instance to another server, if the current server is running short on resources or requires maintenance operations, such as operating system patches. You can use the same technique to relocate Oracle RAC One Node instances to high-capacity servers (for example, to accommodate changes in workload), depending on the resources available in the cluster. In addition, Resource Manager Instance Caging or memory optimization parameters can be set dynamically to further optimize the placement of the Oracle RAC One Node instance on the new server.

By using the Single Client Access Name (SCAN) to connect to the database, clients can locate the service independently of the node on which it is running. Relocating an Oracle RAC One Node instance is, therefore, mostly transparent to the client, depending on the client connection.

For policy-managed Oracle RAC One Node databases, you must ensure that the server pools are configured such that a server will be available for the database to fail over to in case its current node becomes unavailable. In this case, the destination node for online database relocation must be located in the server pool in which the database is located. Alternatively, you can use a server pool of size 1 (one server in the server pool), setting the minimum size to 1 and the importance high enough in relation to all other server pools used in the cluster, to ensure that, upon failure of the one server used in the server pool, a new server from another server pool or the Free server pool is relocated into the server pool, as required.

Oracle RAC One Node and Oracle Clusterware

You can create Oracle RAC One Node databases by using the Database Configuration Assistant (DBCA), as with any other Oracle database (manually created scripts are also a valid alternative). Oracle RAC One Node databases may also be the result of a conversion from either a single-instance Oracle database (using rconfig, for example) or an Oracle RAC database. Typically, Oracle-provided tools register the Oracle RAC One Node database with Oracle Clusterware.

For Oracle RAC One Node databases, you must configure at least one dynamic database service (in addition to and opposed to the default database service). When using an administrator-managed Oracle RAC One Node database, service registration is performed as with any other Oracle RAC database. When you add services to a policy-managed Oracle RAC One Node database, SRVCTL does not accept any placement information, but instead configures those services using the value of the SERVER_POOLS attribute.

Oracle RAC and Network Connectivity

All nodes in an Oracle RAC environment must connect to at least one Local Area Network (LAN) (commonly referred to as the public network) to enable users and applications to access the database. In addition to the public network, Oracle RAC requires private network connectivity used exclusively for communication between the nodes and database instances running on those nodes. This network is commonly referred to as the interconnect.

The interconnect network is a private network that connects all of the servers in the cluster. The interconnect network must use at least one switch and a Gigabit Ethernet adapter. You must configure User Datagram Protocol (UDP) for the cluster interconnect, except in a Windows cluster. Windows clusters use the TCP protocol. On Linux and UNIX systems, you can configure Oracle RAC to use either the UDP or Reliable Data Socket (RDS) protocols for inter-instance communication on the interconnect. Oracle Clusterware uses the same interconnect using the UDP protocol, but cannot be configured to use RDS.

An additional network connectivity is required when using Network Attached Storage (NAS). NAS can be typical NAS devices, such as NFS filers, or can be storage that is connected by using Fibre Channel over IP, for example. This additional network communication channel should be independent of the other communication channels used by Oracle RAC (the public and private network communication). If the storage network communication needs to be converged with one of the other network communication channels, you must ensure that storage-related communication gets first priority.

Benefits of Using RAC

  • High availability: Surviving node and instance failures
  • Scalability: Adding more nodes as you need them in the future
  • Pay as you grow: Paying for only what you need today
  • Key grid computing features: – Growth and shrinkage on demand – Single-button addition of servers – Automatic workload management for services

Oracle Real Application Clusters (RAC) enables high utilization of a cluster of standard, low-cost modular servers such as blades.

RAC offers automatic workload management for services. Services are groups or classifications of applications that comprise business components corresponding to application workloads. Services in RAC enable continuous, uninterrupted database operations and provide support for multiple services on multiple instances. You assign services to run on one or more instances, and alternate instances can serve as backup instances. If a primary instance fails, then Clusterware moves the services from the failed instance to a surviving alternate instance. The Oracle server also automatically load-balances connections across instances hosting a service.

RAC harnesses the power of multiple low-cost computers to serve as a single large computer for database processing, and provides the only viable alternative to large-scale symmetric multiprocessing (SMP) for all types of applications.

RAC, which is based on a shared-disk architecture, can grow and shrink on demand without the need to artificially partition data among the servers of your cluster. RAC also offers a single-button addition of servers to a cluster. Thus, you can easily provide or remove a server to or from the database.

Clusters and Scalability

If your application scales transparently on SMP machines, then it is realistic to expect it to scale well on RAC, without having to make any changes to the application code.

RAC eliminates the database instance, and the node itself, as a single point of failure, and ensures database integrity in the case of such failures.

The following are some scalability examples:

  • Allow more simultaneous batch processes
  • Allow larger degrees of parallelism and more parallel executions to occur
  • Allow large increases in the number of connected users in online transaction processing (OLTP) systems

With Oracle RAC, you can build a cluster that fits your needs, whether the cluster is made up of servers where each server is a two-CPU commodity server or clusters where the servers have 32 or 64 CPUs in each server. The Oracle parallel execution feature allows a single SQL statement to be divided up into multiple processes, where each process completes a subset of work. In an Oracle RAC environment, you can define the parallel processes to run only on the instance where the user is connected to run across multiple instances in the cluster.

Levels of Scalability

Successful implementation of cluster databases requires optimal scalability on four levels:

  • Hardware scalability: Interconnectivity is the key to hardware scalability, which greatly depends on high bandwidth and low latency.
  • Operating system scalability: Methods of synchronization in the operating system can determine the scalability of the system. In some cases, potential scalability of the hardware is lost because of the operating system’s inability to handle multiple resource requests simultaneously.
  • Database management system scalability: A key factor in parallel architectures is whether the parallelism is affected internally or by external processes. The answer to this question affects the synchronization mechanism.
  • Application scalability: Applications must be specifically designed to be scalable. A bottleneck occurs in systems in which every session is updating the same data most of the time. Note that this is not RAC-specific and is true on single-instance systems, too.

It is important to remember that if any of the preceding areas are not scalable (no matter how scalable the other areas are), then parallel cluster processing may not be successful. A typical cause for the lack of scalability is one common shared resource that must be accessed often.

This causes the otherwise parallel operations to serialize on this bottleneck. High latency in the synchronization increases the cost of synchronization, thereby counteracting the benefits of parallelization. This is a general limitation and not a RAC-specific limitation.

Scaleup and Speedup

Scaleup is the ability to sustain the same performance levels (response time) when both workload and resources increase proportionally:

Scaleup = (volume parallel) / (volume original)

For example, if 30 users consume close to 100 percent of the CPU during normal processing, then adding more users would cause the system to slow down due to contention for limited CPU cycles. However, by adding CPUs, you can support extra users without degrading performance.

Speedup is the effect of applying an increasing number of resources to a fixed amount of work to achieve a proportional reduction in execution times:

Speedup = (time original) / (time parallel)

Speedup results in resource availability for other tasks. For example, if queries usually take 10 minutes to process, and running in parallel reduces the time to 5 minutes, then additional queries can run without introducing the contention that might occur if they were to run concurrently.

Speedup/Scaleup and Workloads

The type of workload determines whether scaleup or speedup capabilities can be achieved using parallel processing. Online transaction processing (OLTP) and Internet application environments are characterized by short transactions that cannot be further broken down and, therefore, no speedup can be achieved. However, by deploying greater amounts of resources, a larger volume of transactions can be supported without compromising the response.

Decision support systems (DSS) and parallel query options can attain speedup, as well as scaleup, because they essentially support large tasks without conflicting demands on resources. The parallel query capability within the Oracle database can also be leveraged to decrease overall processing time of long-running queries and to increase the number of such queries that can be run concurrently.

In an environment with a mixed workload of DSS, OLTP, and reporting applications, scaleup can be achieved by running different programs on different hardware. Speedup is possible in a batch environment, but may involve rewriting programs to use the parallel processing capabilities.

Necessity of Global Resources

In single-instance environments, locking coordinates access to a common resource such as a row in a table. Locking prevents two processes from changing the same resource (or row) at the same time.

In RAC environments, internode synchronization is critical because it maintains proper coordination between processes on different nodes, preventing them from changing the same resource at the same time. Internode synchronization guarantees that each instance sees the most recent version of a block in its buffer cache.

Additional Memory Requirement for RAC

RAC-specific memory is mostly allocated in the shared pool at SGA creation time. Because blocks may be cached across instances, you must also account for bigger buffer caches. Therefore, when migrating your Oracle Database from single instance to RAC keeping the workload requirements per instance the same as with the single-instance case, about 10% more buffer cache and 15% more shared pool are needed to run on RAC. These values are heuristics, based on RAC sizing experience. However, these values are mostly upper bounds.

If you use the recommended automatic memory management feature as a starting point, then you can reflect these values in your SGA_TARGET initialization parameter. However, consider that memory requirements per instance are reduced when the same user population is distributed over multiple nodes.

Actual resource usage can be monitored by querying the CURRENT_UTILIZATION and MAX_UTILIZATION columns for the Global Cache Services (GCS) and Global Enqueue Services (GES) entries in the V$RESOURCE_LIMIT view of each instance. You can monitor the exact RAC memory resource usage of the shared pool by querying V$SGASTAT as shown below.

SELECT resource_name,
FROM v$resource_limit
WHERE resource_name like 'g%s_%';
SELECT * FROM v$sgastat
WHERE name like 'g_s%' or name like 'KCL%';

Parallel Execution with RAC

By default, in an Oracle RAC environment, a SQL statement executed in parallel can run across all of the nodes in the cluster. For this cross-node or inter-node parallel execution to perform, the interconnection in the Oracle RAC environment must be size appropriately because inter-node parallel execution may result in a lot of interconnect traffic. If the interconnection has a considerably lower bandwidth in comparison to the I/O bandwidth from the server to the storage subsystem, it may be better to restrict the parallel execution to a single node or to a limited number of nodes. Inter-node parallel execution does not scale with an undersized interconnection.

To limit inter-node parallel execution, you can control parallel execution in an Oracle RAC environment using the PARALLEL_FORCE_LOCAL initialization parameter. By setting this parameter to TRUE, the parallel server processes can only execute on the same Oracle RAC node where the SQL statement was started.

In Oracle Real Application Clusters, services are used to limit the number of instances that participate in a parallel SQL operation. The default service includes all available instances. You can create any number of services, each consisting of one or more instances. Parallel execution servers are to be used only on instances that are members of the specified service. When the parameter PARALLEL_DEGREE_POLICY is set to AUTO, Oracle Database automatically decides if a statement should execute in parallel or not and what degree of parallelism (DOP) it should use.

Oracle RAC can also take advantage of In-Memory Parallel Query. In-Memory Parallel Query loads a large amount of data into the database cache on the physical memory of the servers and execute data processing by using multiple CPU resources.

Oracle Database also determines if the statement can be executed immediately or if it is queued until more system resources are available. Finally, Oracle Database decides if the statement can take advantage of the aggregated cluster memory or not.