Oracle RAC
Oracle RAC ( Oracle R eal A pplication C chandelier ) is an additional option of the database management system of the company Oracle . Oracle RAC enables failure safety by having several nodes of a computer network (English cluster ) access the same database and providing database services for client computers . If one of the nodes fails, the others take over its functionality. In addition, Oracle Real Application Cluster can be used as a scaling option : If the capacity of a database server is insufficient, it can be supplemented by an additional computer node. In order to be able to achieve optimal scaling in the Oracle database cluster, however, the application design must be adapted to the requirements of parallel cluster access. The Oracle Real Application Cluster serves as the basis of the Oracle Grid in the database backend.
architecture
When using the Oracle Real Application Cluster , two or more computer nodes are activated in a cluster (computer network) in normal operation and access the same database. Each node operates its own database instance , but it is connected to the database caches of all computers involved in the cluster using cache fusion technology . For the user, it makes almost no difference which node his access actually arrives at.
The advantage: If a computer fails, clients can connect to a remaining computer node immediately and without restart time. In addition, loads can be distributed across all cluster nodes. Long-running operations can also be parallelized across the entire cluster using parallel slave processes. In addition to the increased availability , the issue of scaling is one of the strengths of the Real Application Cluster. Redundant hardware is not only used in the event of a fault, but can - in contrast to failover clusters - be used in normal operation.
Similar to an Oracle Failover Cluster , a private network for the cluster heartbeat and a separate public network are also required here. However, in the Oracle Real Application Cluster, the database files of a shared storage are competitively accessed. This requires additional communication effort between the nodes, which is necessary for the coordination of locks and changed block images. This cluster communication takes place via the private network. In real application cluster environments, it is therefore important to use the fastest possible private cluster interconnect. At least Gigabit Ethernet should be used. Special, often proprietary solutions with higher bandwidth are even better. Examples are memory channels (alpha-based HP clusters), Myrinet (Linux systems), Scalable Coherent Interconnect (SUN), Veritas LLT (various platforms) or HP Hyper-Fabric HMP.
In the meantime there are also manufacturer solutions such as Oracle Exadata, which are tailored to the use of RAC.
Components
Necessary components for the use of an Oracle Real Application Cluster are:
- Shared storage for all participating cluster nodes (shared access to SCSI , SAN or NAS devices from each of the computers in the cluster)
- Use of a cluster file system , ASM or raw devices on the shared storage
- For each cluster node at least two private and two public network interfaces
- Oracle Cluster Ready Services must always be used from version 10g. In addition, a can cluster management software from another manufacturer to be installed.
- From 11g the cluster is managed by the "Oracle Grid". Additional cluster software is not permitted!
- From 19c (12.2.0.3) the RAC feature will no longer be part of Standard Edition 2!
Advantages and disadvantages
Advantages of an Oracle Real Application Cluster
- Takeover of functionality in the event of a server failure without administrative intervention
- Little time required for restart, client reconnect (renewed connection to the database via another database server in the cluster) occurs transparently for the user or the application during SELECT statements and within a few seconds on the replacement node
- In contrast to the failover cluster , the entire hardware can also be used during normal operation.
- Parallelization can be implemented across all cluster nodes
- Dynamic assignment of services possible
- Compensation of a data center loss if the RAC and its shared media were distributed accordingly ("stretched cluster", up to approx. 2000 m)
Disadvantages of an Oracle Real Application Cluster
- The cluster cannot be stretched over approx. 2000 m ( latency problems ). If the data center fails (for example due to a fire), the entire cluster fails. This error case can be caught by using an additional standby database in combination with Oracle Real Application Clusters or by metro clustering .
- No or only limited interception of logical corruption is possible (block corruption, operating errors ). These errors can be intercepted by a time-delayed standby database ( Oracle Dataguard ), data backup and data recovery combined with a roll-forward or, in many cases, by Oracle Flashback .
- Every time an instance fails, the entire database is not available until the GRD (Global Resource Directory) has been restarted and SMON has determined and received all the blocks required for instance recovery. Only from this point on is it possible to access all blocks except those required for recovery. All data are only fully available when the instance recovery is completely finished.
- If data is requested at the same time or in quick succession on several instances of the cluster which are in the same database blocks, then these are transferred between the buffer caches of the instances concerned. In extreme cases, this can mean that the database is only concerned with moving the blocks back and forth.
Management tools
Oracle Real Application Cluster is supported by the following administration tools:
- Database Configuration Assistant (short: DBCA) for the creation and basic configuration of cluster databases with Oracle
- Oracle Enterprise Manager Grid Control: Graphical user interface for managing an Oracle grid including Oracle Real Application Cluster
- Server Control Utility: Command line-oriented tool for managing databases, services and applications in a real application cluster
See also
literature
- Mike Ault, Madhu Tumma, Ranko Mosic: Oracle 10g Grid & Real Application Cluster , Language: English, Hardcover, 680 pages, Rampant Tech Press, Date of Issue: August 28, 2004, ISBN 0-9744355-4-6
- Tim Donar: Tru64 UNIX-Oracle 9i Cluster Quick Reference , Language: English, paperback - 312 pages - Butterworth-Heinemann, Date of publication: April 30, 2002, ISBN 1-55558-272-9
- Matthew Hart, Scott Jesse: Oracle Database 10g High Availability with RAC, Flashback & Data Guard , Language: English, Paperback - 421 pages - Osborne / McGraw-Hill, Published April 1, 2004, ISBN 0-07-225428-9
- Andrea Held; Oracle 10g high availability with RAC, Data Guard and Flashback , m. CD-ROM, language: German, hardcover - Addison-Wesley, Munich, publication date: October 2004, ISBN 3-8273-2163-8
- Andrea Held: Oracle 11g - New Features , Language: German, Hardcover - Hanser-Verlag, Munich, Publication Date: May 2008, ISBN 3-446-41198-4
- Kristien Hens, Michael Loebmann: Creating Highly Available Database Solutions , Language: English, Paperback, 228 pages, Prentice Hall PTR, Date of publication: January 31, 2005, ISBN 0-13-186390-8
- Murali Vallath: Oracle Real Application Clusters , Language: English, paperback - 781 pages - Digital Press, Published April 2004, ISBN 1-55558-288-5
- Larissa Janssen: High-performance database systems: theory and practice , language: German, paperback, 260 pages, color illustrations, Books on Demand GmbH, date of publication: January 2008, ISBN 978-3-8334-9326-3
- Octavian Lascu, Mustafa Mah, Michel Passet, Harald Hammershøi, SeongLul Son, Maciej Przepiórka: Deploying Oracle 10g RAC on AIX V5 with GPFS , Language: English, paperback - 310 pages - IBM, Date of publication: April 2008, ISBN 0-7384-8583 -7
Web links
Individual evidence
- ↑ iX. Retrieved September 2, 2018 .
- ^ Database Licensing Information User Manual. Retrieved April 8, 2019 (American English).