Shulou(Shulou.com)06/01 Report--

Free download address of PDF e-book

Http://www.allitebooks.com/oracle-database-12c-dba-handbook/

1 Overview

The RAC database is highly available and extensible. Each node must have its own Oralce license and RAC license. The cost of achieving high availability needs to be weighed.

1.1 the hardware is configured with at least two nodes with basic redundancy of server components. The built-in redundancy of shared storage reduces the amount of software redundancy that creates ASM disk groups.

1.2 Software configuration from the Oracle10g built-in Cluster Software Service CRS, you must first install Clusterware in the CRS_HOME directory, and then install the database software on ORACLE_HOME, which can be a directory shared by all nodes (windows), or use OCFS Cluster File system version 2.x or above, otherwise each node installs its own copy.

1.3 the network is configured with at least 3 IP and two NICs per node, which are used for public network, node interconnection, and failover virtual addresses respectively.

2 characteristics

The parameter file SPFILE is shared by all nodes in the cluster on the ASM disk group. The parameter values of different instances can be different, all of which are prefixed with *, otherwise the node name is prefixed.

2.2RAC-related initial parameters: INSTANCE_NUMBER unique instance number, INSTANCE_NAME,CLUSTER_DATABASE cluster is the number of true,CLUSTER_DATABASE_INSTANCES instances, ASM this value is 4unique clusters connections specifies that the cluster IPC traffic network 2.3 dynamic view begins with the GV$ and the additional column INST_ID is always 1, whose rows are the node multiples of the V$ view. 1 General data file view vicious datafile repertoire vandalism temp file2 supports cluster view gv$sessiion3 maintenance 3.1 startup: nodes in RAC can be started in any order, first start ASM to mount disk groups, and then start RDBMS to join the cluster. Linux modifiable / etc/oratab automatically starts each instance 3.2 redo log: RAC each instance has its own online redo log, rolls forward all the information in the log, or rolls back uncommitted transactions with undo tablespaces. A surviving instance can detect other instance failures and use the online redo log to commit transactions. If all instances fail, the first instance started will use the online redo logs of all instances in the cluster to perform instance recovery for the database. If media recovery is required, all instances must be closed and 3.3 undo tablespaces must be performed from one instance: each UNDO tablespace must be visible to all nodes in the cluster, and each instance has its own undo tablespace on shared storage. Failover: automatic connection switching can be achieved by correctly configuring the client program through tnsnames.ora, and the response time for data processing continues to be slightly longer. Adjust the RAC node: an instance can be adjusted using the automatic workload repository AWR. By means of Cloud Control12c analysis.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.