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

1. Data Guard architecture diagram

Figure 1.1 Data Guard architecture diagram

Data Guard (hereinafter referred to as DG) consists of five parts, namely, redo data service, role management and switching, application service of standby library, protection mode of main library, and management of DG.

Second, redo the transmission service

2.1 introduction to redo transport service

The DG redo transfer service coordinates the redo data transfer process from the primary database to the standby database. At the same time, the LGWR in the main database writes the redo data to its own online redo log file (ORL); the LNS process transmits the buffered redo log of the SGA of the primary library to the standby database; and the RFS on the standby database receives the data transmitted by LNS and writes it to the standby redo log (SRL) by LGWR.

Note:

1.ORL:Online Redo Logfile: connect redo log files

2.LNS:Log Network Server: responsible for transferring the redo buffer data from the main library SGA to the standby database

3.RFS:Remote File Server: responsible for receiving data transmitted by LNS.

4.SRL:Standby Redo Logfile: alternate redo log file.

2.2 redo the transport service architecture

Figure 2.1 redoing the transport service

DG supports two transmission methods of LNS processes: synchronous method (SYNC) and asynchronous method (ASYNC).

Synchronous transfer does not allow LNS to acknowledge the success of the commit operation until LGWR confirms that the redo data required for transaction recovery has been written to disk at the standby site. This approach protects all data from loss, but may affect the performance of the main library.

In asynchronous transfer, LGWR does not have to wait for a confirmation message from LNS, which hardly affects the performance of the main library. In asynchronous transmission, you can also enable redo data compression.

2.3 automatic interval processing

Figure 2.2 automatic interval processing

A log interval occurs when the LNS process stops transferring redo data to the standby while the primary continues to commit transactions. During the interruption, a special ARCH process continuous ping backup is used on the main library to determine its status. When the communication with the slave database is restored, the ARCH ping process will query the standby control file through the RFS process to determine the last complete log file received by the standby database from the master database, determine which log files are needed to resynchronize the slave database, and transfer the corresponding log files through other ARCH processes.

During the next log switch, LNS attempts to connect to the slave database and starts to transfer the current redo data after success, while ARCH processes the interval in the background. After the application archive log on the repository catches up with the redo record, the application process no longer reads the archive log and instead reads the current SRL (assuming "real-time application" is configured).

III. Application services

Data Guard provides two different ways to apply redo data to a standby database: physical standby and logical standby, respectively.

Physical standby is an exact physical copy of the primary database block by block; logical standby uses logical processes to apply change coordination to the standby database.

Note: at present, only the physical standby has been used, and the logical standby has not been contacted.

IV. Protection mode

There are three protection modes for DG: the highest performance mode (default), the highest available mode, and the maximum protection mode.

Highest performance mode: pay attention to the performance of the main library, despise data protection, and use ASYNC to redo the transfer.

Highest available mode: most emphasis on availability, followed by zero data loss protection, using SYNC to redo transmission. If the primary database fails to receive a confirmation message from the standby library, it disconnects the LNS connection and continues to commit transactions on the primary database.

Maximum protection mode: put data protection in the first place and redo the transfer using SYNC. If the primary library does not receive a confirmation message from the available standby library, the primary library will stop.

V. role management and role transformation

There are two roles of master library and standby library in DG, and there are two kinds of switching between roles: switchover and failover.

VI. The management mode of DG

There are three ways to manage DG, which are SQL*PLUS, DG Broker and OEM.

Reference:

Chapter 1 of the Oracle Data Guard 11g complete reference Manual.

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.