Oracle RAC and Highly Available Clusters

Oracle RAC (Oracle Real Application Clusters) is an additional option of the database management system company Oracle. Oracle RAC enables resilience by multiple nodes of a computer network in the same database access and computer database for client services.
If a node fails, the others take over its functionality.
In addition, Oracle Real Application Clusters as a way of scaling are:
Is the capacity of a database server does not, this by a further nodes are added.
For optimal scaling on the Oracle database cluster to be able to achieve, but must use the design to the demands of competing clusters requests to be adjusted.
The Oracle Real Application Clusters is used as the basis of the Oracle Grid in the database backend.

Architecture
When using the Oracle Real Application Clusters are in normal operation, two or more nodes in a cluster (computer network) is activated and access to the same data to. Each node operates in a separate database instance, but on the cache fusion technology with the database cache on the cluster of all participating machine is running.

To make it to the user almost no difference to what he and his node access actually arrives.
The advantage: If a computer, so clients can immediately and without recovery time remaining on a computer node.
In addition, loads on all cluster nodes are distributed. Cross-end operations can be parallel-Slave processes over the entire cluster will be parallelized.
That is in addition to the increased availability also the theme scaling one of the strengths of the Real Application Clusters.
Redundant hardware is not only used in case of errors, but - unlike in the case of failover clusters - in normal operation can be used.
Similar to an Oracle failover cluster, it is a private network for the cluster heartbeat and one of them separated Public Network needed.
However, the Oracle Real Application Clusters competitively on the database files in a shared Storages accessed.
This requires additional communication between the nodes, to the vote by blocking and changing block images is necessary. This cluster communication takes place via the Private Network.
In Real Application Clusters environments, it is therefore important that as soon as possible private cluster interconnect to use. At least Gigabit Ethernet should be used. Better still are specific, often proprietary solutions with higher bandwidth.
Examples are

memory channel (alphabasierte HP Cluster),
Myrinet (Linux systems),
Scalable Coherent-interconnect (SUN),
Veritas LLT (various platforms) or HP Hyper-Fabric HMP.

Components

Necessary components for the use of Oracle Real Application Clusters are:

Shared storage for all cluster nodes (shared access to SCSI, SAN or NAS device from any computer in the cluster) Using a clustered file system, ASM or raw devices on the Shared Storage
For each cluster node at least two private and two public-Network Interface
Oracle Cluster Ready Services 10g version must always be used.
In addition, a cluster management software from a third manufacturer to be installed.

Pros and Cons

Advantages of Oracle Real Application Clusters:

Adoption of the functionality in the event of failure of a server without administrative intervention.
Little time needed for recovery, client-reconnect (re-connect to the database on another database server in the cluster) is for the user or the application during SELECT statements are transparent and within a few seconds on the replacement node.
The complete hardware can - unlike the failover cluster - even during normal operation can be used.
Parallelization on all cluster nodes across feasible. Dynamic allocation of services possible
Compensation of a data loss, when the RAC and its shared media was distributed accordingly ( "Stretched cluster", up to about 2000 m)

Disadvantages of an Oracle Real Application Clusters:

Stretching the cluster has about 2000 m no longer possible (latency problems).
In case of failure of the computer center (for example by a fire), the entire cluster. This error event may be adjusted through the use of an additional standby database in combination with Oracle Real Application Clusters or by Metro-clustering be intercepted. No or only limited intercepting logical corruption possible (block corruption, user error). These errors can be delayed by a standby database (Oracle Data Guard), backup and recovery combined with a roll-forward or even in many cases by Oracle Flashback be intercepted.
Each time an instance fails, the entire database is not available, and until such time as the GRD (Global Resource Directory) has been restarted and all SMON for instance recovery necessary blocks identified and also has gotten. Only at this point, it will be possible on all units except those designed for the recovery needs to be accessed. And only when the body recovery is completed, all data is fully available.
Be at the same time on several instances in the cluster data requested, in the same database blocks are located, they will be shared between pools concerned instances transferred. In extreme cases it can cause the database is only engaged in the blocks back and forth to push. The more instances, the greater this effort.

Tools for managing

Oracle Real Application Clusters Administration is determined by the following tools:

Database Configuration Assistant (abbreviated DBCA) to create and base configuration of the cluster with Oracle databases Oracle Enterprise Manager
Grid Control: Graphical user interface for managing an Oracle grid, including Oracle Real Application Clusters Utility
Server Control Utility: Command-based tool for the management of databases, services and applications in a Real Application Clusters

Oracle Errors