Three-minute reading of Oracle data storage disaster Dataguard

Oracle database is still at the forefront of commercial databases. The security of the core business and core data running on the Oracle database is particularly important.

There are generally two broad categories of disaster recovery products that are common to Oracle databases on the market.

• Oracle's own disaster-tolerant products
• Disaster-tolerant products for non-Oracle Companies

Oracle's current disaster recovery products include our common Dataguard and Oracle GoldenGate (OGG) products belonging to the middleware division. Non-Oracle company has DSG Diesche and DDS Nine bridge, these two products and ogg in the implementation principle of the same.

Products such as Oracle Goldengate and DSG/DDS are relatively expensive license, and Dataguard use is free after the acquisition of Oracle RDBMS software license. Oracle 11g Active DataGuard (ADG) requires additional charges.

For data migration, OGG/DDS/DSG benefits are more pronounced. Data Warehouse data centralization to ORALCE data analysis use Dataguard is too bloated, not flexible. Ogg also has significant advantages in synchronizing data migrations on heterogeneous platforms. It is also used for data migration scenarios where downtime is less than the time. The downtime of the business migration can be greatly reduced through its initialization data and later incremental synchronization. Ogg's disaster recovery synchronization can satisfy a variety of needs, such as synchronizing data from only one or several users of a database. While Ogg can implement multi-process extraction, OGG 12.x supports extracting data from the Dataguard repository to the offsite disaster recovery database. Now the more recommended architecture is the Adg+ogg 12.x from the standby library to extract data to the remote disaster recovery room.

Dataguard is superior to Oracle GoldenGate for both maintenance and budget support for Oracle-to-Oracle disaster resilience. Oracle 11g Active DataGuard (ADG) ready-to-read queries can support Reporting Services and distribute pressure from the main library to the standby. The 11g snapshot Dataguard temporarily transforms the standby into a read-write test environment that is fully consistent with production and provides the most realistic simulation environment for testing. Once the test is finished, a key is cut back to the standby mode. The test data is erased automatically, without affecting the data synchronization disaster.

Insert here to say a little bit about Oracle's extended RAC disaster-tolerant high availability. Extended RAC is the traditional RAC two nodes from the same machine room to be placed in two different rooms to achieve single-point disaster recovery. However, consider some of the problems and storage requirements of the RAC itself. RAC two nodes cannot be placed too far (as the network latency increases, the performance of the RAC becomes exponentially weaker, which is why Exadata puts all the RAC nodes in the same cabinet), and the distance is too close to achieve real disaster recovery, such as power grids, typhoons, Floods, earthquakes and so on. Support for storage replication is also required, and the budget will be a big problem.

DataGuard How long does it take to switch the high availability of disaster tolerant? In our existing customer environment, Dataguard is a disaster preparedness environment for Oracle RAC. A one-click Failover script can be used to switch to the standby when a catastrophic damage to the RAC environment is completely unusable. The database itself can be completed in 2 minutes with normal synchronization.

The comparison of Dataguard and OGG on network bandwidth. Dataguard to transmit more data than Ogg, its bandwidth occupies a bit more than Ogg. But it's not as scary as we think. In the normal Dataguard synchronization, Oracle chooses to push the redo data from redo buffer to the standby database via the LNS. These are actually very small quantities, and the requirements for network bandwidth are not very high. The size of the specific amount can be roughly analyzed by the load profile in the AWR report between peak business hours. The ARCH/LGWR attribute has been canceled in the Oracle 11g configuration Dataguard log_archive_dest_n property. The LGWR is used by default for transmission. in other words, unless there is a gap, Oracle dataguard disaster recovery essentially no longer relies on archived logs.

Oracle 11g DataGuard is also much more convenient to implement than 10g. By copying the database duplicate from Active database backup, you can implement the Dataguard environment directly without stopping the main library (provided that the main library has an archive mode turned on, or it requires an outage operation).

If you need to use ADG as a report library, we recommend that you deploy Dataguard monitoring at the same time. A script that monitors the state of ADG synchronization is implemented through the Linux Shell + Oracle procedure. Can monitor whether the real-time synchronization is normal and the delay time.

Three-minute reading of Oracle data storage disaster Dataguard